WO2020175915A1 - Method for encoding/decoding video signal, and apparatus therefor - Google Patents

Abstract

A method for decoding a video, according to the present invention, may comprise the steps of: parsing a first flag indicating whether inter prediction on the basis of a merge mode is applied to a current block; if the first flag is true, parsing a second flag indicating whether a regular merge mode or a merge offset encoding mode is applied to the current block; and if the second flag is true, parsing a third flag indicating whether the merge offset encoding mode is applied to the current block.

Description

2020/175915 1»（：1/10公020/002754 Specification

Name of the invention: Video signal encoding/decoding method and device technology field therefor

[1] The present invention relates to a video signal encoding/decoding method and an apparatus therefor. Background

[2] As the display panel becomes more and more large,

Video services are in demand. The biggest problem with high-definition video services is that the amount of data increases significantly, and to solve this problem, research to improve the video compression rate is actively being conducted.

In 2009, MPEG (Motion Picture Experts Group) and ITU-T (International

The Video Coding Experts Group (VCEG) under the Telecommunication Union-Telecommunication) formed the Joint Collaborative Team on Video Coding (JCT-VC). JCT-VC proposed HEVC (High Efficiency Video Coding), a video compression standard that has about twice the compression performance of H.264/AVC, and was approved as a standard on January 25, 2013. Rapid development of high-definition video services Depending on the situation, the performance of HEVC is gradually pushing its limits.

Detailed description of the invention

Technical task

[3] In the present invention, in encoding/decoding a video signal, the motion information table is

It is an object of the present invention to provide a used post-merge induction method and an apparatus for performing the above method.

[4] An object of the present invention is to provide a method for updating motion information of blocks included in a merge processing area to a motion information table and an apparatus for performing the method in encoding/decoding a video signal.

[5] The present invention is based on the last candidate in encoding/decoding video signals.

It is an object of the present invention to provide a method for refining the induced motion vector and an apparatus for performing the method.

[6] The present invention aims to provide a method for efficiently determining an inter prediction method to be applied to a current block and an apparatus for performing the method in encoding/decoding a video signal.

[7] The technical tasks to be achieved in the present invention are not limited to the technical tasks mentioned above, and other technical tasks that are not mentioned are clear to those with ordinary knowledge in the technical field to which the present invention belongs from the following description. It can be understood.

Problem solving means

[8] The video signal decoding method according to the present invention is based on the current block edge mode. 2020/175915 1»（：1^1{2020/002754 Parsing the first flag indicating whether or not inter prediction is applied, if the first flag is true, the current block is encoded in a regular merge mode or a merge offset. Parsing a second flag indicating whether a mode is applied, and if the second flag is true, parsing a third flag indicating whether the merge offset encoding mode is applied to the current block. have. In this case, when the third flag is true, the merge offset encoding mode is applied to the current block, and when the third flag is false, the regular merge mode may be applied to the current block.

[9] The video signal encoding method according to the present invention includes encoding a first flag indicating whether inter prediction based on a merge mode is applied to the current block, and when the first flag is true, the current block is regular Encoding a second flag indicating whether a merge mode or a merge offset encoding mode is applied, and when the second flag is true, a third flag indicating whether the merge offset encoding mode is applied to the current block It may include the step of encoding. In this case, when the merge offset encoding mode is applied to the current block, the third flag is set to true, and when the regular merge mode is applied to the current block, the third flag may be set to false.

[1] The video signal decoding/encoding method according to the present invention may further include a step of parsing/encoding a fourth flag indicating whether the combined prediction mode is applied to the current block when the second flag is false. have.

[11] In the video signal decoding/encoding method according to the present invention, the prediction unit partitioning-based encoding method can be applied when the fourth flag is false.

[12] In the video signal decoding/encoding method according to the present invention, the motion information of the current block is derived from the merge candidate list of the current block, and the number of merge candidates derived from neighboring blocks of the current block If it is less than the threshold value, motion information candidates included in the motion information table may be added to the merge candidate list as merge candidates.

[13] In the video signal decoding/encoding method according to the present invention, when the current block is included in the merge processing area, the motion information table may not be updated while the blocks included in the merge processing area are decoded.

[14] In the video signal decoding/encoding method according to the present invention, when the current block is included in the merge processing area, based on the position of the current block in the merge processing area, the motion information of the current block is Whether to update the motion information table can be determined.

[15] In the video signal decoding/encoding method according to the present invention, when the current block is located in the lower right corner of the merge processing area, it may be decided to update the motion information of the current block in the motion information table.

[16] The features briefly summarized above for the present invention are 2020/175915 1»（：1^1{2020/002754 This is only an exemplary aspect of the detailed description, and does not limit the scope of the present invention.

Effects of the Invention

[17] According to the present invention, it is possible to improve the inter prediction efficiency by inducing a merge candidate using a motion information table.

[18] According to the present invention, it is possible to improve inter prediction efficiency by providing a method of updating motion information of blocks included in a merge processing area in a motion information table.

[19] According to the present invention, the inter prediction efficiency can be improved by refining the motion vector induced based on the merge candidate.

[2] According to the present invention, it is possible to efficiently determine an inter prediction method to be applied to the current block.

[21] The effects that can be obtained from the present invention are not limited to the effects mentioned above, and other effects that are not mentioned are from the following description.

It will be clearly understood by those with ordinary knowledge in the technical field.

Brief description of the drawing

1 is a block diagram of an image encoder (encoder) according to an embodiment of the present invention.

2 is a block diagram of an image decoder (decoder) according to an embodiment of the present invention.

3 is a diagram showing a basic coding tree unit according to an embodiment of the present invention.

[25] Fig. 4 is a diagram showing various division types of a coding block.

5 is a view showing an example of the division of the coding tree unit.

6 is a flow diagram of an inter prediction method according to an embodiment of the present invention.

[28] Fig. 7 shows a process of inducing movement information of the current block in the merge mode.

It is a flow chart.

[29] FIG. 8 is a diagram illustrating candidate blocks used to induce a merge candidate.

[3] Fig. 9 is a view showing the positions of reference samples.

[31] FIG. 10 is a diagram illustrating candidate blocks used to induce a merge candidate.

11 is a view showing an example in which the position of the reference sample is changed.

12 is a view showing an example in which the position of the reference sample is changed.

13 is a diagram for explaining an update mode of a motion information table.

14 is a diagram showing an update pattern of a motion information table.

15 is a diagram showing an example in which the index of the previously stored motion information candidate is updated.

16 is a diagram showing the location of a representative sub-block.

17 shows an example in which a motion information table is generated for each inter prediction mode.

18 shows an example in which a motion information table is generated for each motion vector resolution.

[4] Fig. 19 shows the motion information of the block to which the merge offset coding method is applied. 2020/175915 1»（：1^1{2020/002754 This is an example that is stored in the information table.

20 is a diagram showing an example in which motion information candidates included in the long-term motion information table are added to the merge candidate list.

21 is a diagram showing an example in which a redundancy test is performed only for some of the merge candidates.

22 is a diagram showing an example in which a redundancy test with a specific merge candidate is omitted.

[44] Fig. 23 shows a candidate block emerged in the same merge processing area as the current block.

It is a drawing showing an example of being set as unavailable as a candidate.

[45] FIG. 24 is a diagram showing an example of inducing a merge candidate for the current block when the current block is included in the merge processing area.

25 is a diagram showing a temporary motion information table.

[47] Fig. 26 is a diagram showing an example of merging the motion information table and the temporary motion information table.

27 is a diagram showing an example of dividing a coding block into a plurality of prediction units using a diagonal line.

28 is a diagram showing an example of dividing a coding block into two prediction units.

[5] Fig. 29 shows examples of dividing a coding block into a plurality of prediction blocks having different sizes.

Fig. 30 is a diagram showing neighboring blocks used to induce a split mode merge candidate.

31 is a diagram for explaining an example of determining availability of neighboring blocks for each prediction unit.

[53] Figs. 32 and 33 are diagrams showing an example in which a predicted sample is derived based on a weighted sum operation of the first predicted sample and the second predicted sample.

방향을 나타내는 (1뇨_{6 011}_：1(1 의값에따른오프셋벡터를나타낸도면이다. [54] Figure 34 shows the size of the offset vector

It is a diagram showing the offset vector according to the value of (1 urine _{6 011} _: 1 (1) indicating the direction.

방향을 나타내는 (1뇨_{6 011}_：1(1 의값에따른오프셋벡터를나타낸도면이다. [55] Figure 35 shows the size of the offset vector

It is a diagram showing the offset vector according to the value of (1 urine _{6 011} _: 1 (1) indicating the direction.

Modes for the implementation of the invention

Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

[57] Encoding and decoding of an image is performed in block units. For example, encoding/decoding processing such as transformation, quantization, prediction, in-loop filtering, or restoration is performed for a coding block, a transformation block, or a prediction block. Can be

[58] Hereinafter, a block to be encoded/decoded will be referred to as a “current block.” As an example, the current block may represent a coding block, a transform block, or a block according to the current encoding/decoding process step.

[59] In addition, the term'unit' used in this specification refers to a specific government encryption/decryption process. 2020/175915 1»（：1^1{2020/002754 It represents the basic unit to be executed, and'block' can be understood as representing a sample array of a predetermined size. Unless otherwise stated,'block' and 'Unit' may be used with equivalent meaning. For example, in an embodiment described below, it may be understood that a coding block and a coding unit have the same meaning.

[6] Fig. 1 is a block diagram of an image encoder (encoder) according to an embodiment of the present invention.

Referring to FIG. 1, the image encoding apparatus 100 includes a picture division unit 110, a prediction unit 120, 125, a conversion unit 130, a quantization unit 135, a re-alignment unit 160, and an entropy. Encoding unit 165,

An inverse quantization unit 140, an inverse transform unit 145, a filter unit 150, and a memory 155 may be included.

[62] Each of the components shown in FIG. 1 is characterized by different characteristics in the image encoding apparatus.

As shown independently to represent functions, it does not mean that each component is made up of separate hardware or a single software component; that is, each component is listed and included in each component for convenience of explanation, and at least two of each component are included. The constituent parts of the present invention are combined to form one constituent part, or one constituent part can be divided into a plurality of constituent parts to perform functions, and the integrated and separate embodiments of each of these constituent parts are within the scope of the present invention as long as they do not depart from the essence of the present invention Included.

[63] In addition, some of the components may not be essential components to perform the essential functions of the present invention, but only optional components to improve performance. The present invention can be implemented by including only the components essential for realizing the essence of the present invention, excluding components used for improving performance, and a structural diagram including only essential components excluding optional components used for performance improvement. It is included in the scope of the invention.

[64] The picture divider (no) can divide the input picture into at least one processing unit. At this time, the processing unit may be a prediction unit (prediction unit:

It may be a unit (Transform Unit: TU) or a coding unit (CU). In the picture division part (no), a picture is divided into a combination of a plurality of coding units, prediction units, and transformation units. A picture can be encoded by selecting a combination of a coding unit, a prediction unit, and a conversion unit based on a predetermined criterion (for example, a cost function).

[65] For example, a picture can be divided into a plurality of coding units. To divide the coding units in a picture, a recursive tree structure such as a quad tree structure can be used. A coding unit that is divided into other coding units by using the largest coding unit as a root can be divided with as many child nodes as the number of divided coding units. A coding unit that is no longer divided according to certain restrictions becomes a leaf node, i.e., assuming that only square division is possible for one coding unit, one coding unit can be divided into up to four different coding units.

[66] Hereinafter, in the embodiment of the present invention, the coding unit is the unit that performs the coding. 2020/175915 1»（：1^1{2020/002754 It can be used as a meaning, or it can be used as a unit that performs decryption.

[67] A prediction unit is at least one of the same size within a coding unit.

It may be divided into a shape such as a square or a rectangle, and one of the prediction units divided within one coding unit may be divided so that one prediction unit has a different shape and/or size from the other prediction unit. have.

[68] When generating a prediction unit that performs intra prediction based on the coding unit

In the case of non-coding units, intra example can be performed without dividing into a plurality of prediction units NxN.

[69] The prediction units 120 and 125 may include an inter prediction unit 120 that performs inter prediction and an intra prediction unit 125 that performs intra prediction. For the prediction unit, whether to use inter prediction or intra prediction You can decide whether to perform the prediction, and determine specific information (e.g., intra prediction mode, motion vector, reference picture, etc.) for each prediction method. At this time, the processing unit in which the prediction is performed, the prediction method, and the specific content are determined. The processing unit may be different. For example, the prediction method and the prediction mode are determined by the prediction unit, and the execution of the prediction may be performed by the transformation unit. The residual value (residual block) between the generated prediction block and the original block can be input to the transform unit 130. In addition, the prediction mode information and motion vector information used for prediction are included in the entropy encoding unit together with the residual value. In case of using a specific encoding mode, it is possible to encode the original block as it is and transmit it to the decoder without generating a prediction block through the prediction units (120, 125).

The P inter prediction unit 120 may predict the prediction unit based on information of at least one of the previous or subsequent pictures of the current picture, and in some cases, based on the information of some regions in which encoding is completed in the current picture. The prediction unit may be predicted by using the inter prediction unit 120. The inter prediction unit 120 may include a reference picture interpolation unit, a motion prediction unit, and a motion compensation unit.

1] The reference picture interpolation unit receives reference picture information from the memory 155 and can generate pixel information less than an integer pixel from the reference picture. In the case of a luminance pixel, pixel information less than an integer pixel is stored in units of 1/4 pixel. To generate, a DCT-based interpolation filter (DCT-based Interpolation Filter) that differs in filter coefficients can be used. In the case of a color difference signal, filter coefficients are generated to generate pixel information less than an integer pixel in 1/8 pixel units. A different DCT-based 4-tap interpolation filter (DCT-based Interpolation Filter) can be used.

2] Motion prediction unit performs motion based on the reference picture interpolated by the reference picture interpolation unit.

Prediction can be performed Various methods such as FBMA (Full search-based Block Matching Algorithm), TSS (Three Step Search), NTS (New Three-Step Search Algorithm) can be used to calculate the motion vector. A vector may have a motion vector value of 1/2 or 1/4 pixel unit based on the interpolated pixel. 2020/175915 1»（：1^1{2020/002754 In the motion prediction section, the current prediction unit can be predicted by different motion prediction methods. Skip method, Merge method, AMVP( Various methods such as Advanced Motion Vector Prediction method and Intra Block Copy method can be used.

3] The intra prediction unit 125 may generate a prediction unit based on reference pixel information around the current block, which is pixel information in the current picture. Since the surrounding block of the current prediction unit is a block on which inter prediction has been performed, the reference pixel is In the case of a pixel subjected to inter prediction, the reference pixel included in the block on which the inter prediction was performed can be used by replacing the reference pixel information of the block on which the intra prediction was performed. The reference pixel information that is not available can be used by replacing at least one reference pixel among the available reference pixels.

4] In intra prediction, the prediction mode may have a directional prediction mode that uses reference pixel information according to the prediction direction, and a non-directional mode that does not use directional information when performing prediction. A mode for predicting luminance information and color difference information are included. The mode for prediction may be different, and the intra prediction mode information or the predicted luminance signal information used to predict the luminance information to predict the color difference information can be utilized.

5] When performing intra prediction, if the size of the prediction unit and the size of the transformation unit are the same, the prediction unit is based on the pixel existing on the left, the pixel on the top left, and the pixel on the top of the prediction unit. Intra prediction can be performed.

However, if the size of the prediction unit and the size of the conversion unit are different when performing intra prediction, intra prediction can be performed using a reference pixel based on the conversion unit. In addition, the NxN division is used only for the smallest coding unit. Intra prediction can be used.

6] Intra prediction method is AIS (Adaptive Intra) in the reference pixel according to the prediction mode.

The prediction block can be created after applying the smoothing) filter. The type of AIS filter applied to the reference pixel may be different. In order to perform the intra prediction method, the intra prediction mode of the current prediction unit exists around the current prediction unit. The prediction can be made from the intra prediction mode of the prediction unit. When the prediction mode of the current prediction unit is predicted by using the mode information predicted from the surrounding prediction unit, a predetermined flag is displayed if the intra prediction mode of the current prediction unit and the peripheral prediction unit are the same. Using the information, information indicating that the prediction mode of the current prediction unit and the surrounding prediction unit is the same can be transmitted, and if the prediction modes of the current prediction unit and the surrounding prediction unit are different, entropy encoding is performed to encode the prediction mode information of the current block. 7] In addition, a residual block containing residual information, which is a difference between the prediction unit and the original block of the prediction unit, based on the prediction unit generated by the prediction units (120, 125). The generated residual block may be input to the transform unit 130. 8] In the transform unit 130, the original block and the prediction unit generated through the prediction units 120 and 125 2020/175915 1»（：1/10公020/002754 The residual block containing residual information can be transformed using a transformation method such as DCT (Discrete Cosine Transform) or DST (Discrete Sine Transform). Here, the DCT conversion core includes at least one of DCT2 or DCT8, and the DST conversion core includes DST7. Whether to apply DCT or DST to convert the residual block, the prediction unit used to generate the residual block Intra prediction mode information of can be determined based on the information of the intra prediction mode of the. Transformation for the residual block can also be skipped. A flag indicating whether to skip the transformation for the residual block can be encoded. Transformation skipping is the size of which is less than or equal to a threshold. May be acceptable for residual blocks, luma components, or chroma components under the 4:4:4 format.

The quantization unit 135 may quantize values converted into the frequency domain by the conversion unit 130. The quantization coefficient may vary depending on the block or the importance of the image. The value calculated by the quantization unit 135 may be provided to the inverse quantization unit 140 and the re-alignment unit 160.

The rearrangement unit 160 may rearrange the coefficient values with respect to the quantized residual values. The rearrangement unit 160 may change the two-dimensional block shape coefficient into a one-dimensional vector shape through a coefficient scanning method. For example,

The rearrangement unit 160 can scan from DC coefficients to coefficients in the high frequency domain and change them to a one-dimensional vector form using a Zig-Zag Scan method. Instead of a zag scan, a vertical scan that scans two-dimensional block shape coefficients in the column direction, and a horizontal scan that scans two-dimensional block shape coefficients in the row direction can also be used, i.e., depending on the size of the transformation unit and the intra prediction mode. You can decide which scan method will be used: zig-zag scan, vertical scan, and horizontal scan.

]]]] The entropy encoding unit 165 uses the values calculated by the re-alignment unit 160 as the basis.

24 o511

78888891 Entropy coding can be performed. Entropy coding is for example exponential

Various coding methods such as is (Exponential Golomb), CAVLC (Context-Adaptive Variable Length Coding), and CABAC (Context-Adaptive Binary Arithmetic Coding) can be used.

83] The entropy encoding unit 165 is the residual value coefficient information and block type information, prediction mode information, division unit information, prediction unit information and transmission unit information of the coding unit from the reordering unit 160 and the prediction units 120 and 125. , Motion vector information, reference frame information, block interpolation information, filtering information, etc. can be encoded.

The entropy encoding unit 165 may entropy-encode the coefficient value of the encoding unit inputted from the reordering unit 160.

The inverse quantization unit 140 and the inverse transform unit 145 inverse quantize values quantized in the quantization unit 135 and inverse transform the values transformed by the transform unit _130. Inverse quantization unit 140 and inverse transform unit 145 ), the residual value (Residual) generated in the prediction unit (120, 125) is combined with the prediction unit predicted through the motion estimation unit, motion compensation unit, and intra prediction unit. 2020/175915 1»（：1/10公020/002754 Reconstructed Block) can be created.

The filter unit 150 may include at least one of a deblocking filter, an offset correction unit, and an adaptive loop filter (ALF).

The deblocking filter can remove block distortion caused by the boundary between blocks in the restored picture. To determine whether to perform deblocking, the current block is based on the pixels included in several columns or rows in the block. It is possible to judge whether to apply a deblocking filter. When applying a deblocking filter to a block, a strong filter or a weak filter can be applied depending on the required deblocking filtering strength. Also, a deblocking filter can be applied. In doing so, when performing vertical filtering and horizontal filtering, horizontal direction filtering and vertical direction filtering can be processed in parallel.

88] The offset correction unit can correct the offset from the original image on a pixel-by-pixel basis for the deblocking image. After dividing the pixels included in the image into a certain number of areas to perform offset correction for a specific picture, A method of determining an area to perform the offset and applying the offset to the area, or a method of applying the offset in consideration of edge information of each pixel can be used.

Adaptive Loop Filtering (ALF) can be performed on the basis of a value obtained by comparing the filtered restored image and the original image. After dividing the pixels included in the image into predetermined groups, one filter to be applied to the group is determined for each group. Filtering can be performed differentially. Information related to whether or not to apply ALF, the luminance signal can be transmitted for each coding unit (CU), and the shape and filter coefficient of the ALF filter to be applied may vary according to each block. In addition, the characteristics of the block to be applied may be different. The same type (fixed type) of ALF filter may be applied regardless of.

[9]] The memory 155 can store the restored block or picture calculated through the filter unit 150.

9211

99 9 8887461, and the stored restored block or picture may be provided to the prediction units 120 and 125 when performing inter prediction.

2 is a block diagram of an image decoder (decoder) according to an embodiment of the present invention. 2, the image decoder 200 is an entropy decoding unit 210,

A rearrangement unit 215, an inverse quantization unit 220, an inverse transform unit 225, prediction units 230 and 235, a filter unit 240, and a memory 245 may be included.

[93] When a video bitstream is input from a video encoder, the input bitstream can be decoded in a procedure opposite to that of the video encoder.

The entropy decoding unit (2W) can perform entropy decoding in a procedure opposite to that of entropy encoding performed by the entropy encoding unit of the image encoder. For example, Exponential Golomb (Exponential Golomb) in response to the method performed by the image encoder ), CAVLC(Context- Adaptive Variable Length Coding),

Various methods such as CAB AC (Context-Adaptive Binary Arithmetic Coding) can be applied. 2020/175915 1»（：1^1{2020/002754

[95] In the entropy decoding unit 210, intra prediction and inter prediction performed by the encoder

Information related to prediction can be decoded.

[96] The rearrangement unit 215 is the entropy decryption unit 210

Reordering can be performed based on the method of rearranging the bitstream in the encoder.

The coefficients expressed in the form of a one-dimensional vector can be restored and rearranged into coefficients in the form of a two-dimensional block. The re-alignment unit 215 receives information related to the coefficient scanning performed in the encoding unit, and the scanning performed in the encoding unit Reordering can be done through the reverse scanning method based on the sequence.

[97] The inverse quantization unit 220 may perform inverse quantization based on the quantization parameter provided by the encoder and the coefficient value of the re-aligned block.

다.또는,영상부호화기에서변환이 스킵된경우,역변환부 (225)에서도역변환을수행하지않을수있다.역변환은 영상부호화기에서결정된전송단위를기초로수행될수있다.영상복호화기의 역변환부 (225)에서는예측방법,현재블록의크기및예측방향등복수의 정보에따라변환기법 (예를들어,。(그또는 0 )이선택적으로수행될수있다. [98] The inverse transformation unit 225 includes at least one of 2 or 0018 to perform an inverse transformation, that is, an inverse 1) (that or an inverse example!) in the transformation unit with respect to the quantization result performed by the image encoder, and

C. Alternatively, if the transformation is skipped in the image encoder, the inverse transformation may not be performed in the inverse transformation unit 225. The inverse transformation may be performed based on the transmission unit determined in the image encoder. The inverse transformation unit 225 of the image decoder In the case, a transformation technique (for example, .(or 0)) can be selectively performed according to a plurality of pieces of information such as the prediction method, the size of the current block, and the prediction direction.

The prediction units 230 and 235 may generate a block based on information related to prediction block generation provided by the entropy decoding unit 210 and previously decoded block or picture information provided from the memory 245, for example.

[100] As described above, if the size of the prediction unit and the size of the transformation unit are the same when the intra prediction is performed in the same manner as the operation of the image encoder, the pixel existing on the left side of the prediction unit, the pixel existing on the top left, and the top Intra prediction is performed for the prediction unit based on the pixels existing in the prediction unit, but when the size of the prediction unit and the size of the conversion unit are different when performing intra prediction, intra prediction is performed using a reference pixel based on the conversion unit. You can also use intra prediction that uses an NxN division for only the smallest coding unit.

[101] The prediction units 230 and 235 may include a prediction unit determination unit, an inter prediction unit, and an intra prediction unit. The prediction unit determination unit information about prediction units input from the entropy decoding unit 210, prediction of an intra prediction method. By receiving various information such as mode information and motion prediction related information of the inter prediction method, it is possible to classify the prediction unit in the current coding unit, and to determine whether the prediction unit performs inter prediction or intra prediction. (230) uses the information necessary for inter prediction of the current prediction unit provided by the image encoder, based on the information contained in at least one picture of the current picture containing the current prediction unit or the picture after the current prediction unit. Inter-prediction can be performed, or inter-prediction can be performed based on the information of some regions previously restored within the current picture containing the current prediction unit. 2020/175915 1»（：1^1{2020/002754 There is.

[102] In order to perform inter prediction, the coding unit is

The motion prediction method of the included prediction unit can be judged whether the method is skip mode, merge mode (Merge mode), motion vector prediction mode (AMVP mode), or intrablock copy mode.

[103] The intra prediction unit 235 may generate a prediction block based on pixel information in the current picture. When the prediction unit is a prediction unit that has performed intra prediction, an image

The intra prediction mode information of the prediction unit provided by the encoder can be used to perform intra prediction mode information. The intra example name section 235 may include an AIS (Adaptive Intra Smoothing) filter, a reference pixel interpolation section, and a DC filter. The AIS filter is a part that performs filtering on the reference pixel of the current block, and can be applied by determining whether to apply the filter according to the prediction mode of the current prediction unit. Using the prediction mode of the prediction unit and AIS filter information provided by the image encoder AIS filtering can be performed on the reference pixel of the current block. When the prediction mode of the current block is a mode that does not perform AIS filtering, the AIS filter may not be applied.

[104] When the prediction mode of the prediction unit is a prediction unit that performs intra prediction based on the pixel value obtained by interpolating the reference pixel, the reference pixel interpolation unit can generate a reference pixel of a pixel unit less than an integer value by interpolating the reference pixel. If the prediction mode of the current prediction unit is a prediction mode that generates a prediction block without interpolating a reference pixel, the reference pixel may not be interpolated. The DC filter can generate a prediction block through filtering when the prediction mode of the current block is DC mode.

[105] The reconstructed block or picture may be provided to the filter unit 240. The filter unit 240

It can include a deblocking filter, an offset correction unit, and ALF.

[106] Information on whether the deblocking filter is applied to the block or picture from the video encoder, and when the deblocking filter is applied, a strong filter is applied.

You can be provided with information on whether you have applied or a weak filter has been applied. In the deblocking filter of the image decoder, information related to the deblocking filter provided by the image encoder is provided, and the image decoder can perform deblocking filtering on the corresponding block.

[107] The offset correction unit can perform offset correction on the restored image based on the type of offset correction applied to the image at the time of encoding and information on the offset value.

[108] ALF is the ALF application information and ALF coefficient information provided from the encoder.

It can be applied to a coding unit as a basis. This ALF information can be provided by being included in a specific parameter set.

[109] The memory 245 stores the restored picture or block and stores the reference picture or reference.

It can be made available as a block, and the restored picture can be provided as an output unit.

[11 this 2020/175915 1»（：1^1{2020/002754

3 is a diagram showing a basic coding tree unit according to an embodiment of the present invention.

[112] A coding block of the largest size can be defined as a coding tree block. One picture is divided into a plurality of coding tree units (CTU). The coding tree unit is a coding unit of the largest size, and LCU (LCU). Largest Coding Unit). FIG. 3 shows an example in which one picture is divided into a plurality of coding tree units.

[113] The size of the coding tree unit can be defined at the picture level or the sequence level. To this end, information indicating the size of the coding tree unit can be signaled through a picture parameter set or a sequence parameter set.

[114] For example, the size of the coding tree unit for all pictures in the sequence is 128x128.

Alternatively, either 128x128 or 256x256 at the picture level can be determined as the size of the coding tree unit. For example, the size of the coding tree unit is set to 128x128 in the first picture, and the coding tree is set to 128x128 in the second picture. Unit size

It can be set to 256x256.

[115] By dividing the coding tree unit, it is possible to generate a coding block.

Represents a basic unit for encoding/decoding processing. For example, prediction or transformation may be performed for each coding block, or a predictive encoding mode may be determined for each coding block. Here, the predictive encoding mode refers to the method of generating the predictive image. For example, the encoding mode is an intra-screen example (Intra Prediction, Intra example), and inter-screen

For example (Inter Prediction, i.e.), current picture reference (Current Picture Referencing, CPR, or Intra Block Copy (IBC)) or Combined Yes (Combined)

Prediction) For a coding block, at least one of intra prediction, inter prediction, current picture reference, or combined prediction can be used to generate a prediction block for the coding block.

[116] Information indicating the prediction coding mode of the current block is transmitted through the bitstream.

For example, the information may be a 1-bit flag indicating whether the predictive encoding mode is intra mode or inter mode. Only when the predictive encoding mode of the current block is determined to be inter mode, the current picture reference or Combined prediction may be available.

[117] Reference current picture sets the current picture as the reference picture, and

This is to obtain the prediction block of the current block from the region where the coding/decoding is completed. Here, the current picture means a picture including the current block. Information indicating whether or not the current picture reference is applied to the current block is the bitstream. For example, the information may be a 1-bit flag. If the flag is true, the predictive encoding mode of the current block is referred to as the current picture reference.

Is determined, and if the flag is false, the prediction mode of the current block may be determined by inter prediction.

[118] Alternatively, based on the reference picture index, the predictive encoding mode of the current block can be determined. 2020/175915 1»（：1^1{2020/002754. For example, when the reference picture index points to the current picture, the prediction coding mode of the current block can be determined by the current picture reference. The reference picture index is the current picture reference. When pointing to a picture other than the other, the prediction encoding mode of the current block can be determined by inter prediction, i.e., the current picture reference is a prediction method using information of the area in which the current picture has been internally encoded/decoded, and the inter prediction is the encoding/decoding. This is a prediction method using information of another picture that has been decoded.

[119] Combined prediction is a combination of two or more of intra prediction, inter prediction, and current picture reference.

For example, when combined prediction is applied, a first prediction block may be generated based on one of intra prediction, inter prediction, or current picture reference, and a second prediction block may be generated based on the other. When the first prediction block and the second prediction block are generated, a final prediction block may be generated through an average operation or a weighting operation of the first prediction block and the second prediction block. Information indicating whether the combined prediction is applied is provided in the bitstream. It can be signaled through. The information can be a 1-bit flag.

4 is a diagram showing various division types of a coding block.

[121] Coding block is divided into quadtree, binary tree, or triple tree.

A divided coding block can be divided into a plurality of coding blocks based on a quad tree division, a binary tree division, or a triple tree division.

[122] Quadtree partitioning represents a partitioning technique that divides the current block into four blocks. As a result of quadtree division, the current block can be divided into 4 square partitions (see Fig. 4)-!, reference).

[123] Binary tree division is a division technique that divides the current block into two blocks.

방향 바이너리트리분할결과를나타낸것이다. Dividing the current block into two blocks along the vertical direction (i.e., using a vertical line crossing the current block) can be called a vertical binary tree division, and along the horizontal direction (i.e., the current block horizontally). Dividing the current block into two blocks can be referred to as a horizontal direction binary tree division. As a result of the binary tree division, the current block can be divided into two amorphous partitions. Binary tree splitting result

This is the result of dividing the direction binary tree.

[124] Triple tree division is a division technique that divides the current block into 3 blocks.

Dividing the current block into three blocks along the vertical direction (i.e., using two vertical lines crossing the current block) can be referred to as a vertical triple tree division, and along the horizontal direction (i.e., the current block Dividing the current block into three blocks can be referred to as horizontal triple-tree division. As a result of the triple-tree division, the current block can be divided into three non-square partitions. At this time, the partition located in the center of the current block 2020/175915 1»（：1^1{2020/002754 The width/height can be twice the width/height of other partitions. In Fig. 4(d)'SPLIT_TT_VER' shows the result of the vertical triple tree division, (E)'SPLIT_TT_HOR' in 4 shows the result of splitting a triple tree in the horizontal direction.

[125] The number of divisions of the coding tree unit can be defined as the partitioning depth. The maximum division depth of the coding tree unit can be determined at the sequence or picture level. Accordingly, the maximum division depth of the coding tree unit may be different for each sequence or field.

[126] Alternatively, the maximum dividing depth for each of the dividing techniques can be determined individually. For example, the maximum partition depth allowed for a quadtree partition may be different from the maximum partition depth allowed for a binary tree partition and/or a triple tree partition.

[127] The encoder can signal information representing at least one of the partition shape or the partition depth of the current block through the bitstream.

Based on the information parsed from the bitstream, it is possible to determine the division type and division depth of the coding tree unit.

5 is a diagram showing an example of a division of a coding tree unit.

[129] Divisions such as quadtree division, binary tree division, and/or triple tree division

Dividing the coding block using the multi-tree division

Partitioning).

[130] Coding blocks generated by applying multi-tree division to a coding block can be referred to as sub-coding blocks. When the division depth of a coding block is no, the division depth of sub-coding blocks is set to k+1.

[131] Conversely, for coding blocks with a split depth of k+1, an encoding block with a split depth can be referred to as an upper coding block.

[132] The partition type of the current coding block is the partition type of the upper coding block or neighbor coding.

It can be determined based on at least one of the partition types of the block, where the neighboring coding block is adjacent to the current coding block, and one of the top neighboring blocks of the current coding block, left neighboring blocks, or neighboring blocks adjacent to the upper left corner. Here, the partition type may include at least one of whether to divide a quadtree, whether to divide a binary tree, a direction to divide a binary tree, whether to divide a triple tree, or a direction to divide a triple tree.

[133] In order to determine the division type of the coding block, whether or not the coding block is divided

The indicated information can be signaled through the bitstream. The information is a 1-bit flag'split_cu_flag', and the fact that the flag is true indicates that the coding block is divided by the head tree division technique.

[134] When split_cu_flag is true, information indicating whether the coding block is divided into quadtrees may be signaled through the bitstream. The information is a 1-bit flag split_qt_flag, and when the flag is true, the coding block is 4 It can be divided into blocks. 2020/175915 1»（：1^1{2020/002754

As an example, in the example shown in FIG. 5, as the coding tree unit is divided into a quad tree, it is shown that four coding blocks having a division depth of 1 are generated. In addition, 4 generated as a result of the quad tree division. It is shown that quadtree division is again applied to the first and fourth coding blocks among the four coding blocks. As a result, four coding blocks with a division depth of 2 can be generated.

[136] In addition, by applying the quadtree division again to a coding block having a division depth of 2, a coding block having a division depth of 3 can be generated.

[137] When quad-tree division is not applied to a coding block, considering at least one of the size of the coding block, whether the coding block is located at the picture boundary, the maximum division depth, or the shape of the neighboring block, It is possible to determine whether to perform binary tree division or triple tree division. When it is determined that binary tree division or triple tree division is performed in the coding block, information indicating the division direction can be signaled through the bitstream. The information may be a 1-bit lag mtt_split_cu_vertical_flag. Based on the lag, whether the dividing direction is a vertical direction or a horizontal direction can be determined. In addition, either binary tree division or triple tree division is the coding. Information indicating whether the block is applied can be signaled through the bitstream. The information may be a 1-bit flag mtt_split_cu_binary_flag. Based on the flag, it may be determined whether binary re-tree division or triple-tree division is applied to the coding block.

[138] As an example, in the example shown in FIG. 5, a vertical direction is applied to an encoding block with a division depth of 1

It is shown that binary tree division is applied, vertical triple tree division is applied to the left coding block among the coding blocks generated as a result of the division, and vertical binary tree division is applied to the right coding block.

[139]

[140] Inter prediction is a prediction coding mode that predicts the current block using information of the previous picture. For example, a block (hereinafter, a collocated block) at the same position as the current block in the previous picture is currently It can be set as a prediction block of a block. Hereinafter, a prediction block generated based on a block at the same position as the current block will be referred to as a collocated, i.e., a block (Collocated Prediction Block).

[141] On the other hand, an object existing in the previous picture is moved to a different position in the current picture.

If it has moved, the current block can be effectively predicted by using the motion of the object. For example, if the moving direction and size of the object can be known by comparing the previous picture with the current picture, the current block's prediction is made taking into account the motion information of the object. Blocks (or prediction images) can be created. Hereinafter, a prediction block generated using motion information may be referred to as a motion prediction block.

[142] A residual block can be generated by differentiating the prediction block from the current block. In this case, if the object movement exists, the collocated prediction block 2020/175915 1»（：1^1{2020/002754 By using the motion prediction block instead, the energy of the residual block can be reduced, and accordingly, the compression performance of the residual block can be improved.

[143] As above, motion compensation is performed to generate prediction blocks using motion information.

It can be referred to as prediction. In most inter prediction, a prediction block can be generated based on motion compensation prediction.

[144] The motion information may include at least one of a motion vector, a reference picture index, a predictive direction or a bidirectional weight index. The motion vector indicates the moving direction and size of the object. The reference picture index is a reference picture included in the reference picture list. Among the pictures, the reference picture of the current block is specified. The prediction direction indicates either one-way L0 prediction, one-way L1 prediction, or two-way prediction (L0 prediction and L1 prediction). Depending on the prediction direction of the current block, movement in the L0 direction. At least one of human information or motion information in the L1 direction can be used. The bidirectional weight index specifies the weight applied to the L0 prediction block and the weight applied to the L1, i.e., the block.

6 is a flow chart of an inter prediction method according to an embodiment of the present invention.

Referring to FIG. 6, the inter prediction method includes: determining an inter prediction mode of a current block (S601), acquiring motion information of a current block according to the determined inter prediction mode (S602), and acquired movement Based on the information, a step of performing motion compensation prediction for the current block (S603).

[147] Here, the inter prediction mode represents various techniques for determining the motion information of the current block, and an inter prediction mode using translation motion information and an inter prediction mode using affine motion information. For example, the inter prediction mode using translational motion information includes a merge mode and a motion vector prediction mode, and the inter prediction mode using the affine motion information includes an afine merge mode and an afine motion vector prediction mode. The motion information of the current block may be determined based on information parsed from a neighboring block or a bitstream adjacent to the current block according to the inter prediction mode.

[148] The motion information of the current block may be derived from the motion information of the other block of the current block. Here, the other block may be a block encoded/decoded by inter prediction prior to the current block. The motion information of the current block can be derived from the motion information of the other block. Setting the same as the motion information can be defined as the merge mode. In addition, setting the motion vector of the other block as the predicted value of the motion vector of the current block can be defined as a motion vector prediction mode.

7 shows a process of inducing motion information of a current block in a merge mode.

It is a flow chart.

[150] The merge candidate of the current block can be derived (S701). The merge candidate of the current block can be derived from the block encoded/decoded by inter prediction before the current block.

8 is a diagram illustrating candidate blocks used to induce a merge candidate.

[152] Candidate blocks are neighboring blocks containing samples adjacent to the current block or 2020/175915 1»（：1^1{2020/002754 May contain at least one of the non-neighboring blocks containing samples that are not adjacent to the block. Hereinafter, the samples that determine candidate blocks are defined as reference samples. . In addition, a reference sample adjacent to the current block will be referred to as a neighbor reference sample, and a reference sample not adjacent to the current block will be referred to as a non-neighbor reference sample.

[153] The neighboring reference sample may be included in the neighboring column of the leftmost column of the current block or the neighboring row of the topmost row of the current block. For example, when the coordinates of the upper left sample of the current block are (0, 0). , Blocks containing the reference samples at positions (-1, 11-1), Blocks containing the reference samples at positions (-1, -1), Blocks containing reference samples at positions (, -1), (- At least one of the blocks containing the reference sample at positions 1 and ¾ or the block containing the reference sample at positions (-1, -1) can be used as a candidate block. Referring to the drawing, the neighbors of index 0 to index 4 can be used. Blocks can be used as candidate blocks.

[154] The non-neighboring reference sample represents a sample in which at least one of the X-axis distance or axial distance from the reference sample adjacent to the current block has a predefined value. For example, the X-axis distance from the left reference sample is predefined. A block containing a reference sample that is a specified value, a block containing a non-neighbor sample whose X-axis distance from the upper reference sample is a predefined value, or a non-neighbor sample whose X-axis distance and the X-axis distance from the left upper reference sample are predefined values. At least one of the blocks containing the can be used as a candidate block. The predefined value can be a natural number such as 4, 8, 12, 16, etc. Referring to the drawing, at least one of the blocks of index 5 to 26 can be used as a candidate block. It can be used as a block.

[155] Not located on the same vertical, horizontal or diagonal line as the neighboring reference sample.

You can also set the sample as a non-neighbor-based sample.

9 is a diagram showing the positions of reference samples.

인경우,상단이웃기준샘플로부터 축으로 N만큼 떨어진상단비이웃기준샘플의위치는 (( /2)-1, -1-비으로설정되고,상단이웃 기준샘플로부터 축으로 2N만큼떨어진상단비이웃기준샘플의위치는 (0, -1-2비으로설정될수있다.즉,비인접기준샘플의위치는인접기준샘플의 위치와인접기준샘플과의거리를기초로결정될수있다. As in the example shown in FIG. 9, the X coordinates of the upper non-neighboring reference samples may be set differently from the X coordinates of the upper neighboring reference samples. For example, the position of the upper neighboring reference sample is

In the case of, the position of the top non-neighbor reference sample, which is N away from the top neighbor reference sample, is set to (( /2)-1, -1-ratio, and the top non-neighbor is 2N away from the top neighbor reference sample. The position of the reference sample can be set in the ratio of (0, -1-2, i.e., the position of the non-adjacent reference sample can be determined based on the position of the adjacent reference sample and the distance from the adjacent reference sample.

[158] Hereinafter, among the candidate blocks, a candidate block including a neighbor reference sample is referred to as a neighbor block, and a block including a non-neighbor reference sample is referred to as a neighbor block.

[159] When the distance between the current block and the candidate block is greater than or equal to the threshold value, the candidate block may be set to be unavailable as a merge candidate. The threshold value may be determined based on the size of the coding tree unit. For example, The threshold value is the

, 설정될수있다.오프셋 N는부호화기및 복호화기에서기정의된값으로, 4, 8, 16,

있다. 2020/175915 1»（：1^1{2020/002754 By adding or subtracting the offset from the height of the unit

, Can be set Offset N is a value predefined in the encoder and decoder, 4, 8, 16,

have. 2020/175915 1»（：1^1{2020/002754

[160] If the difference between the X-axis coordinate of the current block and the X-axis coordinate of the sample included in the candidate block is greater than the threshold value, the candidate block may be determined to be unavailable as a future candidate.

[161] Or, candidate blocks that do not belong to the same coding tree unit as the current block are merged.

As an example, if the reference sample deviates from the upper boundary of the coding tree unit to which the current block belongs, the candidate block including the reference sample may be set to be unavailable as a future candidate.

[162] If the upper boundary of the current block is adjacent to the upper boundary of the coding tree unit, it is determined that a number of candidate blocks are not usable as a merge candidate, and the encoding/decoding efficiency of the current block may decrease. To solve the same problem, candidate blocks can be set so that the number of candidate blocks positioned at the left of the current block is greater than the number of candidate blocks positioned at the top of the current block.

10 is a diagram illustrating candidate blocks used to induce a merge candidate.

[164] As in the example shown in FIG. 1, the upper blocks belonging to the upper N block rows of the current block and the left blocks belonging to the left block rows of the current block can be set as candidate blocks. In this case, M is more than N. By setting it larger, the number of left candidate blocks can be set larger than the number of upper candidate blocks.

[165] For example, it is possible to set the difference between the X-axis coordinate of the reference sample in the current block and the X-axis coordinate of the upper block that can be used as a candidate block not to exceed a multiple of the current block height. In addition, the reference sample in the current block It can be set so that the difference between the X-axis coordinate of and the X-axis coordinate of the left block that can be used as a candidate block does not exceed M times the current block width.

[166] As an example, in the example shown in FIG.

Blocks and blocks belonging to the five block columns to the left of the current block are shown to be set as candidate blocks.

[167] As another example, the candidate block does not belong to the same coding tree unit as the current block.

In this case, instead of the candidate block, a block belonging to the same coding tree unit as the current block or a block including a reference sample adjacent to the boundary of the coding tree unit may be used to induce a merge candidate.

11 is a view showing an example in which the position of the reference sample is changed.

[169] When a reference sample is included in a coding tree unit different from the current block, and the reference sample does not close to the boundary of the coding tree unit, a reference sample adjacent to the boundary of the coding tree unit is used instead of the reference sample. Can determine a candidate block.

[17 is an example, in Fig. 11] and (in the non-shown example, the upper boundary and coding of the current block)

When the upper boundary of the tree unit is in contact, the reference samples at the top of the current block belong to a different coding tree unit from the current block. Among the reference samples belonging to a coding tree unit different from the current block, the upper boundary of the coding tree unit is not adjacent to the current block. 2020/175915 1»（：1^1{2020/002754 The reference sample can be replaced with the sample adjacent to the upper boundary of the coding tree unit.

[171] As an example, as in the example shown in Fig. 11), the reference sample at position 6 is replaced with the sample at position 6'located at the upper boundary of the coding tree unit, and As in the previous example, the reference sample at position 15 can be replaced with a sample at position 15' located at the upper boundary of the coding tree unit. At this time, the X coordinate of the replacement sample is changed to the adjacent position of the coding tree unit, The X coordinate of the replacement sample can be set the same as the reference sample. For example, the sample at position 6'has the same X coordinate as the sample at position 6, and the sample at position 15' has the same X coordinate as the sample at position 15. Can have

[172] Alternatively, the X coordinate of the replacement sample by adding or subtracting the offset from the X coordinate of the reference sample can be set. For example, the X coordinate of the neighboring reference sample and the non-neighboring reference sample located at the top of the current block In the same case, an offset can be added or subtracted from the reference sample's X coordinate as the replacement sample's X coordinate, which means that the replacement sample replacing the non-neighboring reference sample has the same position as the other non-neighboring reference sample or the neighboring reference sample. It is to prevent it from becoming.

12 is a view showing an example in which the position of the reference sample is changed.

[174] In replacing a reference sample that is included in a coding tree unit different from the current block and is not adjacent to the boundary of the coding tree unit with a sample located at the boundary of the coding tree unit, an offset is added from the X coordinate of the reference sample or Subtracted value of the replacement sample

It can be set in the X coordinate.

차분한값으로설정될 수있다. [175] As an example, in the example shown in FIG. 12, the reference sample at position 6 and the reference sample at position 15 are the sample at the 6'position and 15' in which the X coordinate is the same as the row adjacent to the upper boundary of the coding tree unit. In this case, the X coordinate of the sample at position 6'is set as the difference \¥/2 from the X coordinate of the reference sample at position 6, and the X coordinate of the sample at position 15' is at 15 position. In the X coordinate of the reference sample of

It can be set to a subdued value.

[176] Unlike the examples shown in Figs. 11 and 12, at the top of the top row of the current block

It is also possible to set the X coordinate of the row to be located or the X coordinate of the upper boundary of the coding tree unit as the X coordinate of the replacement sample.

[177] Although not shown, it is also possible to determine a sample to replace the reference sample based on the left boundary of the coding tree unit. For example, the reference sample is not included in the same coding tree unit as the current block, and the left boundary of the coding tree unit. If it is not adjacent to, the reference sample can be replaced with a sample that is adjacent to the left boundary of the Coding Tree unit. In this case, the replacement sample has the same coordinates as the reference sample, or by adding or subtracting an offset from the X coordinate of the reference sample. It can have the X-coordinate obtained by doing it.

[178] Thereafter, a block including the replacement sample is set as a candidate block, and a merge candidate of the current block can be derived based on the candidate block.

[179] It is also possible to derive a merge candidate from a temporal neighbor block included in a picture different from the current block. For example, a collocated picture included in a collocated picture. 2020/175915 1»（：1^1{2020/002754 Merge candidates can be derived from the block. Any one of the reference pictures included in the reference picture processing list can be set as a collocated picture. Among the reference pictures, a collo Index information for identifying the cate picture may be signaled through the bitstream. Or, among the reference pictures, a reference picture having a predefined index

It can be determined as a collocated picture.

[180] The motion information of the merge candidate can be set the same as the motion information of the candidate block.

For example, at least one of the motion vector of the candidate block, the reference picture index, the prediction direction, or the bidirectional weight index can be set as the motion information of the merge candidate.

[181] A merge candidate list containing merge candidates can be created (702).

[182] Merge Candidate List The indexes of the merged candidates may be assigned in a predetermined order. For example, a merge candidate derived from the left neighbor block, a merge candidate derived from the upper neighbor block, a merge candidate derived from the upper right neighbor block, a merge candidate derived from the lower left neighbor block, and a merge candidate derived from the upper left neighbor block. And indexes in the order of merge candidates derived from the temporal neighbor block.

있다. [183] When a merge candidate includes multiple candidate candidates, at least one of the plurality of candidate candidates may be selected 703). Specifically, information for specifying any one of the plurality of candidate candidates is signaled through the bitstream. For example, information indicating an index of any one of the merge candidates included in the merge candidate list.

have.

[184]

[185] If the number of merge candidates included in the merge candidate list is less than the threshold value, the motion information candidate included in the motion information table can be added to the merge candidate list as a merge candidate. Here, the threshold value is the merge candidate list. It may be the number of maximum merge candidates that can be included or the number of maximum merge candidates minus the offset. The offset may be a natural number such as 1 or 2.

[186] The motion information table is encoded/decoded based on the inter prediction in the current picture.

Includes motion information candidates derived from the block. For example, motion information

The motion information of the motion information candidates included in the table is based on inter prediction.

It may be set to be the same as the motion information of the encoded/decoded block. Here, the motion information may include at least one of a motion vector, a reference picture index, a prediction direction, or a bidirectional weight index.

[187] Motion information candidates included in the motion information table may be referred to as inter-area merge candidates or predicted-area merge candidates.

[188] The maximum number of motion information candidates that the motion information table can include may be predefined in the encoder and decoder. For example, the maximum number of motion information candidates that the motion information table can contain is 1, 2, 3, 4, 5, 6, 7, 8 or more (e.g.,

16) can be. 2020/175915 1»（：1^1{2020/002754

[189] Or, the maximum number of motion information candidates that the motion information table can contain.

The indicated information may be signaled through a bitstream. The information may be signaled at the sequence, picture, or slice level. The information may indicate the maximum number of motion information candidates that the motion information table can contain. or , The information may represent a difference between the maximum number of motion information candidates that the motion information table can contain and the maximum number of merge candidates that the merge candidate list can contain.

Alternatively, the maximum number of motion information candidates that the motion information table can contain may be determined according to the size of the picture, the size of the slice, or the size of the coding tree unit.

[191] The motion information table can be initialized in units of picture, slice, tile, brick, coding tree unit, or coding tree unit line (row or column). For example, when a slice is initialized, the motion information table is also initialized. , The motion information table may not contain any motion information candidates.

[192] Or, information indicating whether to initialize the motion information table

It can also be signaled through a bitstream. The information can be signaled at the slice, tile, brick or block level. Until the information instructs to initialize the motion information table, the constructed motion information table can be used.

[193] Alternatively, information about the initial motion information candidate may be signaled through a picture parameter set or a slice header. Even if the slice is initialized, the motion information table may contain the initial motion information candidate. Accordingly, the first encoding in the slice /The initial motion information candidate can also be used for the block to be decrypted.

[194] Alternatively, the motion information candidate included in the motion information table of the previous coding tree unit can be set as the initial motion information candidate. For example, the dex among the motion information candidates included in the motion information table of the previous coding tree unit is the most The small motion information candidate or the motion information candidate with the largest index can be set as the initial motion information candidate.

[195] Blocks are encoded/decoded according to the encoding/decoding order, but blocks encoded/decoded based on inter prediction can be sequentially set as motion information candidates according to the encoding/decoding order.

13 is a diagram for explaining an update mode of a motion information table.

[197] For the current block, when inter prediction is performed 1301), motion information candidates can be derived based on the current block (別302). The motion information of the motion information candidate is set the same as the motion information of the current block. Can be

[198] When the motion information table is empty 1303), the motion information candidate derived based on the current block can be added to the motion information table (別304).

[199] When the motion information table already contains motion information candidates 1303), it is possible to perform a redundancy check for the motion information (or motion information candidates derived based on this) of the current block (別305). The inspection checks whether the motion information of the motion information candidate stored in the motion information table and the motion information of the current block are the same. 2020/175915 1»（：1^1{2020/002754 This is to determine. The redundancy check can be performed on all motion information candidates previously stored in the motion information table. Or, motion information candidates previously stored in the motion information table. A redundancy test may be performed for motion information candidates whose dex is above or below a threshold value. Or, a redundancy test may be performed for a predefined number of motion information candidates. For example, 2 with a small index of 2 Two motion information candidates or two Wim motion information candidates with a large index may be determined as a redundancy test object.

[200] A motion information candidate having the same motion information as the motion information of the current block

If not included, motion information candidates derived based on the current block can be added to the motion information table 1308). Whether the motion information candidates are the same is determined by the motion information of the motion information candidates (e.g., motion vector and/or It can be determined based on whether the reference picture index, etc.) are the same.

[201] At this time, if the maximum number of motion information candidates is already stored in the motion information table 1306), the oldest motion information candidate is deleted and 1307), the motion information candidate derived based on the current block is added to the motion information table. Can be added (別308). Here, the oldest motion information candidate may be a motion information candidate with the largest index or a motion information candidate with the smallest index.

[202] Motion information candidates can be identified by respective indexes. When a motion information candidate derived from the current block is added to the motion information table, the lowest index (eg, 0) is assigned to the motion information candidate, and The index of the stored motion information candidates can be increased by 1. At this time, if the maximum number of motion information candidates are already stored in the motion information table, the motion information candidate with the largest index is removed.

[203] Alternatively, the motion information candidate derived from the current block is added to the motion information table.

When added, the largest index can be assigned to the motion information candidate. For example, if the number of motion information candidates previously stored in the motion information table is less than the maximum value, the motion information candidate has the number of previously stored motion information candidates. In addition, if the number of motion information candidates stored in the motion information table is equal to the maximum value, the motion information candidate can be assigned an index subtracting 1 from the maximum value. The motion information candidate with the smallest index is removed, and the remaining indexes of the motion information candidates stored are 1

Decrease.

14 is a diagram showing an update pattern of a motion information table.

[205] It is assumed that the motion information candidate derived from the current block is added to the motion information table, and the largest index is allocated to the motion information candidate. In addition, the motion information table already stores the maximum number of motion information candidates. I assume.

추가하는경우,기 저장된모션정보후보들중인덱스가가장 2020/175915 1»（：1^1{2020/002754 작은모션정보후보 1¾1 ₁幻&11（1 [이를삭제하고,잔여모션정보후보들의 인덱스를 1씩감소시킬수있다.또한,현재블록으로부터유도된모션정보후보 1 &11（1[11+1]의인덱스를최대값（도 14에도시된예에서는미으로설정할수 있다. [206] Motion information candidate 1¾1 ₁幻 1[11+1] derived from the current block is the motion information table

In the case of adding, the previously saved motion information 2020/175915 1»（：1^1{2020/002754 Small motion information candidate 1¾1 ₁幻&11（1 [You can delete this and decrease the index of the residual motion information candidates by one. Also, the motion induced from the current block can be decreased. The index of information candidates 1 & 11 (1[11+1]) can be set to the maximum value (in the example shown in Fig. 14, not).

[207] Going the same motion information candidate as the induced motion information candidate based on the current block

If it is stored 1305), the motion information candidate derived based on the current block may not be added to the motion information table 1309).

[208] Or, the motion information candidate derived based on the current block is added to the motion information table.

In addition, it is also possible to remove the previously stored motion information candidate as the motion information candidate. In this case, the same effect as the previously stored motion information candidate index is updated.

15 is a diagram showing an example in which an index of a previously stored motion information candidate is updated.

[210] Motion information candidates derived based on the current block 111 (If the index of motion information candidates stored in the same time as 111  1 is urine, the previously stored motion information candidates are deleted, and the index is 111 (motion information candidates greater than urine) For example, in the example shown in Fig. 15, 111< 1 is deleted from the same Gamotion information table 1切1111 ¾11（1], and ^1幻 （1[3] to 1¾1 The index up to ^1幻 (1[11] is shown as decreasing by one.

[211] And, motion information candidate derived based on the current block · 0 （1 is motion information

You can add it to the end of the table.

[212] Alternatively, the index assigned to the motion information candidate that is previously stored identical to the motion information candidate derived based on the current block can be updated. For example, the index of the previously stored motion information candidate can be changed to a minimum value or a maximum value.

[213] The motion information of the blocks included in the predetermined area is not added to the motion information table.

For example, motion information candidates derived based on the motion information of blocks included in the merge processing area may not be added to the motion information table. Since the encoding/decoding order of the blocks included in the merge processing area is not defined, it is inappropriate to use one of these motion information for inter prediction of other blocks. Accordingly, the block included in the merge processing area is inappropriate. Motion information candidates derived on the basis of these may not be added to the motion information table.

[214] Alternatively, the motion information of the block smaller than the preset size is displayed in the motion information table.

For example, motion information of a coding block with a width or height less than 4 or 8, or motion information candidates derived based on the motion information of a 4x4 coding block may not be added to the motion information table. have.

[215] When motion compensation prediction is performed on a sub-block basis, motion information candidates can be derived based on the motion information of the representative sub-block among a plurality of sub-blocks included in the current block. For example, for the current block When a sub-block merge candidate is used, the motion information candidate is selected based on the motion information of the representative sub-block among the sub-blocks. 2020/175915 1»（：1^1{2020/002754 Can be derived.

[216] The motion vectors of the subblocks can be derived in the following order: First, one of the merge candidates included in the merge candidate list of the current block is selected, based on the motion vector of the selected merge candidate, and an initial shift vector.知:^ You can induce burrs. Then, the position of the reference sample (e.g., upper left sample or middle position sample) of the coding block is added with the initial shift vector to the position 江315, 8, and the position of the reference sample is 江001815, 0181). Equation 1 below represents the equation for deriving the shift subblock.

[217] [Equation 1]

[218] Then, in the center position of the sub-block containing the ratio, ( 0 1)

The motion vector of the corresponding collocated block can be set as the motion vector of the sub-block including the ratio (31) and 3 ratio.

[219] The representative sub-block can mean a sub-block including the top left sample, the center sample, the bottom right sample, the top right sample, or the bottom left sample of the current block.

[22] Fig. 16 is a diagram showing the location of the representative sub-block.

[221] The example of Fig. 16 shows an example in which the sub-block located at the top left of the current block is set as the representative sub-block, and Figure 16 shows an example in which the sub-block located at the center of the current block is set as the representative sub-block. When motion compensation prediction is performed in units of sub-blocks, motion information candidates for the current block can be derived based on the motion vector of the sub-block containing the upper left sample of the current block or the sub-block containing the center sample of the current block. have.

[222] It is also possible to decide whether to use the current block as a motion information candidate based on the inter prediction mode of the current block. For example, based on the afine motion model.

The coded/decoded block may be set to be unavailable as a motion information candidate. Accordingly, even if the current block is coded/decoded by inter prediction, in the case where the inter prediction mode of the current block is a fine prediction mode, it is based on the current block. It is possible not to update the motion information table.

[223] Alternatively, it is possible to determine whether to use the current block as a motion information candidate based on at least one of the motion vector resolution of the current block, whether the merge offset encoding method is applied, whether the joint prediction is applied, and whether or not the triangular partitioning is applied. For example, if the motion information resolution of the current block is more than 2 integer pels, if the current block has a joint prediction applied, if the current block has triangular partitioning applied, or if the current block has the remaining offset coding method applied, at least one of them, The current block may be set as unavailable as motion information candidates.

[224] Alternatively, motion information candidates may be derived based on at least one subblock vector of the subblocks included in the encoded/decoded block based on the affine motion model. For example, in the upper left of the current block. Sub-block located, sub-block located in the center 2020/175915 1»（：1^1{2020/002754 Or the motion information candidate can be derived by using the subblock located in the upper right corner. Or, the average value of the subblock vectors of the multiple subblocks is calculated as the motion of the motion information candidate. It can also be set as a vector.

[225] Or, the affine seed of the coded/decrypted block based on the afine motion model

It is also possible to induce motion information candidates based on the average value of the vectors. For example, at least one of the first affine seed vector, the second affine seed vector, or the third affine seed vector of the current block is the motion information candidate. It can be set as a motion vector of.

[226] Alternatively, a motion information table can be configured for each inter prediction mode. For example, a motion information table for a block encoded/decoded with an intra block copy, a motion information table for an encoded/decoded block based on a translational motion model. At least one of the motion information tables for the coded/decoded block can be defined based on the table or the affine motion model. Depending on the inter prediction mode of the current block, any one of a plurality of motion information tables can be selected.

17 shows an example in which a motion information table is generated for each inter prediction mode.

[228] When the block is encoded/decoded based on the non-affinity motion model, the block is

Motion information candidate mvCand derived as a basis is a non-agnostic motion information table

It can be added to HmvpCandList. On the other hand, if a block is encoded/decoded based on the afine motion model, the motion information candidate mvAfCand derived based on the block can be added to the afine motion information table HmvpAfCandList.

[229] Motion derived from encoded/decrypted blocks based on afine motion model

The information candidate may store the affine seed vectors of the block. Accordingly, the motion information candidate can be used as a merge candidate for inducing the affine seed vector of the current block.

[23] Or, a motion information table can be configured for each corresponding degree of a motion vector. For example, a motion information table for storing motion information with a resolution of 1/16 pel-in, and a motion information table with a resolution of ¼ pel-in motion vector. Motion information table for saving, motion information table for storing motion information with 1/2 pel-in motion vector resolution, motion information table for storing pel-in motion information with motion vector resolution, or 4 integer felt-in motion with motion vector resolution At least one of the motion information tables for storing information can be defined.

18 shows an example in which a motion information table is generated for each motion vector resolution.

[232] If the motion vector resolution of a block has 1/4 pel, the motion information mvCand of the block can be stored in the quarter-pel motion information table HmvpQPCandList. On the other hand, the motion vector resolution of the block is integer-pel. ), it is possible to store the motion information mvCand of the block in the integer felt motion information table HmvpIPCandList. If the motion vector resolution of the block has 4 integer-pels, the 4 integer felt motion information table HmvpIPCandList Motion information mvCand can be saved. 2020/175915 1»（：1^1{2020/002754

용하여 ,현재 블록의 머지후보를유도할수있다.반면,현재블록의움직임 벡터 해상도가

용하여 ,현재 블록의 머지후보를유도할수있다. [233] According to the motion vector resolution of the current block, it is possible to induce a merge candidate of the current block by selecting the motion information table. For example, the motion vector resolution of the current block is

Can be used to induce a merge candidate for the current block, whereas the motion vector resolution of the current block is

Can be used to induce merge candidates for the current block.

Alternatively, the motion information of the block to which the merge offset encoding method is applied can be stored in a separate motion information table.

19 shows an example in which motion information of a block to which the merge offset encoding method is applied is stored in a separate motion information table.

있다.반면,블록에 머지오프셋벡터부호화방법이 적용된경우,블록의움직임정보마 산를

장하지 않고,머지오프셋모션정보테이블[236] When the merge offset vector encoding method is not applied to the block, the motion information of the block 111< (1 is the motion information table.

On the other hand, when the merge offset vector encoding method is applied to the block, the motion information of the block is calculated.

Without obstacle, merge offset motion information table

It can be stored in HmvpMMVDCandList.

용하여 ,현재블록의 머지후보를유도할수있다.반면,현재블록에머지오프셋부호화방법이 적용된경우,머지오프셋모션정보테이블 HmvpMMVDCandList를이용하여 , 현재블록의머지후보를유도할수있다. [237] Depending on whether the merge offset vector encoding method is applied to the current block, the motion information table can be selected. For example, the merge offset encoding method is used in the current block.

On the other hand, when the merge offset encoding method is applied to the current block, the merge offset motion information table HmvpMMVDCandList can be used to induce the merge candidate of the current block.

[238] In addition to the motion information table described above, it is also possible to define additional motion information tables. In addition to the motion information table described above (hereinafter referred to as the first motion information table), a long term motion information table (hereinafter referred to as the second motion information table) may be defined. Here, the long term motion information table includes long term motion information candidates. do.

[239] When both the first motion information table and the second motion information table are empty, first, a motion information candidate can be added to the second motion information table. The number of motion information candidates available in the second motion information table is limited. After reaching the maximum number, it is possible to add motion information candidates to the first motion information table as a ratio.

[24] Alternatively, one motion information candidate may be added to both the second motion information table and the first motion information table.

[241] At this time, the second motion information table, which has been configured, is no longer updated.

Alternatively, if the decoded area is more than a predetermined ratio of the slice, the second motion information table can be updated, or the second motion information table can be updated every N coding tree unit lines.

[242] On the other hand, the first motion information table can be updated whenever a block encoded/decoded by inter prediction occurs. However, the motion information candidate added to the second motion information table is used to update the first motion information table. May be set not to 2020/175915 1»（：1^1{2020/002754 There is.

[243] Information for selecting either the first motion information table or the second motion information table may be signaled through a bitstream. When the number of merge candidates included in the merge candidate list is less than the threshold, the above The motion information candidates included in the motion information table for which information is indicated can be added to the merge candidate list as merge candidates.

Alternatively, a motion information table may be selected based on the size, shape, inter prediction mode, bi-directional prediction, motion vector refinement, or triangular partitioning of the current block.

[245] Or, even if the motion information candidate included in the first motion information table is added,

If the number of merge candidates included in the candidate list is less than the maximum number, motion information candidates included in the second motion information table can be added to the merge candidate list.

20 is a diagram showing an example in which motion information candidates included in the long term motion information table are added to the merge candidate list.

[247] If the number of merge candidates included in the merge candidate list is less than the maximum number, the motion information candidates included in the first motion information table HmvpCandList can be added to the merge candidate list. The motion information candidates included in the first motion information table Even though information candidates are added to the merge candidate list, if the number of merge candidates included in the merge candidate list is less than the maximum number, motion information candidates included in the long-term motion information table HmvpLTCandList can be added to the merge candidate list.

[248] Table 1 shows the motion information candidates included in the long term motion information table.

It shows the process of adding to the list.

[249] [Table 1]

[25 This motion information candidate can be set to include additional information in addition to motion information. For example, at least one of the size, shape, or partition information of the block can be additionally stored for the motion information candidate. When configuring the merge candidate list of the current block, the current block, size, shape, or partition information among the motion information candidates Use only the same or similar motion information candidate, or if the current block and size, shape, or partition information 2020/175915 1»(：1^1{2020/002754 You can add the same or similar motion information candidates to the merge candidate list first. Or, you can create a motion information table by block size, shape, or partition information. Among the motion information tables, a motion information table corresponding to the shape, size, or partition information of the current block can be used to construct a list of candidates for the current block.

[251] If the number of merge candidates included in the merge candidate list of the current block is less than the threshold value, the motion information candidates included in the motion information table can be added to the merge candidate list as merge candidates. The indexes of the candidates are sorted in ascending or descending order. For example, the motion information candidate with the largest index can be added to the merge candidate list of the current block.

[252] When adding motion information candidates included in the motion information table to the merge candidate list, a redundancy check may be performed between the motion information candidates and the merge candidates previously stored in the merge candidate list. Motion information candidates with the same motion information as the candidate may not be added to the merge candidate list.

[253] As an example, Table 2 shows the process of adding motion information candidates to the merge candidate list.

[254] [Table 2]

[255] Redundancy check is performed on some of the motion information candidates included in the motion information table.

For example, a redundancy test can be performed only for motion information candidates whose index is above or below the threshold value, or N motion information candidates with the largest index or N motion information candidates with the smallest index. A redundancy check can be performed only for the merge candidate list, or a redundancy check can be performed only for some of the merge candidates previously stored in the merge candidate list. For example, a merge candidate whose index is above or below the threshold value or derived from a block at a specific location The redundancy test may only be performed on the old merged candidate. Here, the specific location is the left neighboring block, the upper neighboring block, and the upper right 2020/175915 1»（：1^1{2020/002754 Can contain at least one of the neighboring blocks or the lower left neighboring blocks.

21 is a diagram showing an example in which a redundancy test is performed only for some of the merge candidates

스트에에추가하고자하는경우, 모션정보후보에 대해 인덱스가가장큰 2개의머지후보들 [257] Motion

To add to the stream, the two merge candidates with the largest index for motion information candidates

111 can perform a redundancy test with 60&11(11 1; 11111] 6-2] and 1116 60&11(11 1;[]^11111| 6 6-1], where NumMerge is the available spatial merge candidate and temporal It can indicate the number of merge candidates.

[258] Unlike the example shown, motion information candidates 1¾1 ₁幻 (¾] were added to the merge candidate list.

If you wish to add it, you can also perform a redundancy check for motion information candidates with up to two merge candidates with the smallest index, e.g., when 111 ^(11 [this and whether or not.

서만중복성 검사를수행할수 있다.일 예로,현재블록의좌측에위치하는주변블록으로부터유도된머지 후보또는현재블록의상단에 위치하는주변블록으로부터유도된머지후보중 적어도하나에 대해중복성검사를수행할수있다.머지후보리스트에특정 위치에서유도된머지후보가존재하지 않는경우,중복성 검사없이모션정보 후보를머지후보리스트에추가할수있다. [259]

For example, it is possible to perform a redundancy check on at least one of a merge candidate derived from a neighboring block positioned to the left of the current block or a merge candidate derived from a neighboring block positioned at the top of the current block. If there is no merge candidate derived from a specific location in the merge candidate list, motion information candidates can be added to the merge candidate list without a redundancy check.

스트에에추가하고자하는경우, 모션정보후보에 대해 인덱스가가장큰 2개의머지후보들 [26 this

To add to the stream, the two merge candidates with the largest index for motion information candidates

111 can perform a redundancy test with 60&11(11 1; 11111] 6-2] and 1116 60&11(11 1;[]^11111| 6 6-1], where NumMerge is the available spatial merge candidate and temporal It can indicate the number of merge candidates.

[261] It is also possible to perform a redundancy check with the merge candidate for only some of the motion information candidates. For example, only N motion information candidates with a large index or N motion information candidates with a small index among the motion information candidates included in the motion information table. A redundancy check can be performed. For example, a redundancy check can be performed only for motion information candidates having an index in which the number and difference of motion information candidates included in the motion information table is less than or equal to the threshold value. If the threshold value is 2, motion information can be performed. The redundancy test can be performed only for the three motion information candidates with the largest dex value among the motion information candidates included in the information table. For the motion information candidates excluding the three motion information candidates, the redundancy test may be omitted. If is omitted, the motion information candidate can be added to the merge candidate list, regardless of whether they have the same motion information as the merge candidate.

[262] Conversely, it is also possible to set the redundancy check to be performed only for motion information candidates whose indexes have an index greater than or equal to the threshold value between the number and difference of the motion information candidates included in the motion information table.

[263] The number of motion information candidates for which the redundancy check is performed is determined by the encoder and the decoder. 2020/175915 1»（：1^1{2020/002754 may be predefined. For example, the threshold may be an integer such as 0, 1 or 2.

Alternatively, the threshold may be determined based on at least one of the number of merge candidates included in the merge candidate list or the number of motion information candidates included in the motion information table.

[265] If the same merger candidate as the first motion information candidate is found, the redundancy test between the first motion information candidate and the same merge candidate may be omitted during the redundancy test for the second motion information candidate.

22 is a diagram showing an example in which the redundancy test with a specific merge candidate is omitted.

스트에추가하고자 하는경우,상기모션정보후보와머지후보리스트에기 저장된머지후보들 사이의중복성검사가수행된

동일한

머지 후보리스트에추가하지 않고,인덱스가 1인모션정보후보 1^1幻&11（111-1]와 머지후보들간의중복성검사를수행할수있다.이때,모션정보후보 [267]

When adding to the list, a redundancy check is performed between the motion information candidate and the merge candidates previously stored in the merge candidate list.

same

Without adding to the merge candidate list, a motion information candidate with an index of 1^1幻&11 (111-1) and a redundancy test between merge candidates can be performed. At this time, motion information candidates can perform a redundancy test.

The redundancy test between 1¾11¥1) 0&11（ -1] and merge candidates 1116 & 11（11 1；[]] can be omitted.

동일한 것으로결정되었다.이에따라,

는머지후보리스트에추가되지 않고, &!1 따-1]에 대한중복성 검사가수행될수있다.이때, [268] For example,

It was decided to be the same.

Is not added to the merge candidate list, and a redundancy check for &!1 da-1] may be performed.

13 _\011] The redundancy test between urine 11（1 -1] and 1116 60&11（11 1;[2] may be omitted.

[269] If the number of merge candidates included in the current block's merge candidate list is less than the threshold value, in addition to motion information candidates, at least one of a pairwise merge candidate or a zero merge candidate may be included. It means a merge candidate having the average of the motion vectors of two or more merge candidates as a motion vector, and a zero merge candidate means a merge candidate with a motion vector of 0.

[27 In the merge candidate list of this current block, the merge candidate can be added in the following order.

[271] Spatial Merge Candidate-Temporal Merge Candidate-Motion Information Candidate-(Afine Motion Information Candidate)-Pairwise Merge Candidate-Zero Merge Candidate

[272] Spatial merge candidates are from at least one of the neighboring or non-neighboring blocks.

The affine motion information candidate represents the motion information candidate derived from the block encoded/decoded with the afine motion model, while the derived merge candidate means the temporal merge candidate derived from the previous reference picture.

[273] Motion information tables can also be used in motion vector prediction mode. For example, when the number of motion vector prediction candidates included in the motion vector prediction candidate list of the current block is less than the threshold value, motion included in the motion information table The information candidate can be set as the motion vector prediction candidate for the current block. Specifically, the motion vector of the motion information candidate can be set as the motion vector prediction candidate. 2020/175915 1»（：1^1{2020/002754

[274] When any one of the motion vector prediction candidates included in the motion vector prediction candidate list of the current block is selected, the selected candidate can be set as the motion vector predictor of the current block. Thereafter, the motion vector residual value of the current block is decoded. After that, the motion vector of the current block can be obtained by summing the motion vector predictor and the motion vector residual value.

[275] The motion vector prediction candidate list of the current block can be constructed in the following order.

[276] Spatial motion vector prediction candidate-Temporal motion vector prediction candidate-Motion information candidate-(Afine motion information candidate)-Zero motion vector prediction candidate

[277] The spatial motion vector prediction candidate is at least a neighboring block or a non-neighboring block.

The motion vector prediction candidate derived from one means the candidate, the temporal motion vector prediction candidate means the motion vector prediction candidate derived from the previous reference picture. The affine motion information candidate is derived from the block encoded/decoded with the afine motion model. The zero motion vector prediction candidate represents the candidate whose motion vector value is 0.

[278]

[279] A merge processing area of a larger size than the coding block can be defined.

The coding blocks included in the region are not sequentially encoded/decoded, but can be processed in parallel. Here, that the sequentially encoded/decrypted blocks mean that the encoding/decoding order is not defined. Accordingly, merge processing The encoding/decoding process of blocks included in the region can be independently processed. Alternatively, blocks included in the merge processing area can share merge candidates. Here, the merge candidates can be derived based on the merge processing area.

[28] According to the above-described feature, the merge processing region may be referred to as a parallel processing region, a shared merge region (SMR), or a merge estimation region (MER).

[281] The merge candidate of the current block can be derived based on the coding block. However, if the current block is included in the merge processing area of a larger size than the current block, the candidate block included in the same merge processing area as the current block is It may be set as unavailable as a merge candidate.

[282] Fig. 23 shows a candidate block emerging included in the same merge processing area as the current block.

It is a drawing showing an example of being set as unavailable as a candidate.

[283] In the example shown in (a) of FIG. 23, when encoding/decoding CU5, the reference adjacent to CU5

Blocks containing samples may be set as candidate blocks. At this time, candidate blocks X3 and X4 included in the same merge processing area as CU5 may be set to be unavailable as merge candidates for CU5, while merged same as CU5. Candidate blocks XO, XI and X2 not included in the processing area may be set to be available as merge candidates.

[284] Fig. 23 (in the non-shown example, when the CU8 is encoded/decoded, the reference adjacent to the CU8 2020/175915 1»（：1^1{2020/002754 Blocks containing samples can be set as candidate blocks. At this time, candidate blocks X XI and 8 contained in the merge processing area equal to 18 are ratios as merge candidates. On the other hand, candidate blocks 5 and 9 that are not included in the same merge area as [刀8] can be set as available as merge candidates.

Alternatively, when the current block is included in the merge processing area, a neighboring block adjacent to the current block and a neighboring block adjacent to the merge processing area may be set as candidate blocks.

[286] FIG. 24 is a diagram showing an example of inducing a merge candidate for the current block when the current block is included in the merge processing area.

As in the example shown in FIG. 24, neighboring blocks adjacent to the current block may be set as candidate blocks for inducing merge candidates of the current block. At this time, in the same merge processing area as the current block. The included candidate block may be set as unusable as a merge candidate. For example, when the merge candidate for the coding block 013 is held, the coding block (: top neighbor block 3 and the top right neighbor block included in the merge processing area identical to 113) 4 may be set to be unavailable as a merge candidate for the coding block 0凡.

[288] By scanning neighboring blocks adjacent to the current block according to a predefined order, it is possible to induce a merge candidate. For example, the predefined order may be in the order of, 3, 4, and 2.

[289] Merge candidates that can be derived from neighboring blocks adjacent to the current block

If the number is smaller than the maximum number of the number of merge candidates or the value obtained by subtracting the offset from the maximum number, the merge process for the current block is used using neighboring blocks adjacent to the merge processing area of Fig. 24 (as in the non-shown example). Candidate can be derived. For example, neighboring blocks adjacent to the merge processing area including the coding block can be set as candidate blocks for the coding block (: 3). Here, the neighboring blocks adjacent to the merge processing area can be set. It may include at least one of XI of the left neighboring block, 3 of the upper neighboring blocks, the lower left neighboring block, the upper right neighboring block 4, or the upper left neighboring block.

[29] It is possible to induce merge candidates by scanning the neighboring blocks adjacent to this merge processing area in a predefined order. For example, the predefined order is the order of XI, \3, \4, 0, and 2 Can be

[291] In summary, the merge candidate for the coding block ⑶ 3 included in the merge processing area can be derived by scanning the candidate blocks according to the following scan sequence.

[292] （yl, 3, 0,） _/ 2, XI, 3, 4,, ^2)

[293] However, the scanning order of the candidate blocks illustrated above is only an example of the present invention, and it is possible to scan the candidate blocks in a different order from the above example. Or, the current block or the merge processing area. It is also possible to adaptively determine the scan order based on at least one of the size or shape.

[294] The merge processing area may be square or amorphous. To determine the merge processing area Information for 2020/175915 1»（：1^1{2020/002754 can be signaled through a bitstream. The above information includes at least one of information indicating the shape of the merge processing area or information indicating the size of the merge processing area. When the merge processing area is non-square, at least one of information indicating the size of the merge processing area, information indicating the width and/or height of the merge processing area, or information indicating the ratio between the width and the height of the merge processing area. One can be signaled through the bitstream.

[295] The size of the merge processing area may be determined based on at least one of information signaled through the bitstream, picture resolution, slice size, or tile size.

[296] When motion compensation prediction is performed for a block included in the merge processing area, motion information candidates derived based on the motion information of the block on which the motion compensation prediction was performed can be added to the motion information table.

[297] However, when a motion information candidate derived from a block included in the merge processing area is added to the motion information table, when encoding/decoding other blocks in the merge processing area whose encoding/decoding is actually later than the block, from the block In other words, the motion information of other blocks included in the merge process area may be used even though the dependency between blocks must be eliminated when encoding/decoding blocks included in the merge process area. In order to solve this problem, even if the encoding/decoding of the block included in the merge processing area is completed, the motion information of the block where the encoding/decoding is completed is displayed in the motion information table. May not be added to.

[298] Or, using only the block at the defined position in the merge processing area, motion information

The table can be updated. The predefined positions are the block located in the upper left corner, the block located in the upper right corner, the block located in the lower left corner, the block located in the lower right corner, and the block located in the center. It may include at least one of a block adjacent to the right boundary or a block adjacent to the lower boundary. For example, only the motion information of the block adjacent to the lower right corner in the merge processing area is updated in the motion information table, and the movement of other blocks Information may not be updated in the motion information table.

[299] Or, after decryption of all blocks included in the merge processing area is completed, the

Motion information candidates derived from blocks can be added to the motion information table, that is, the motion information table may not be updated while blocks included in the merge processing area are being encoded/decoded.

[300] As an example, when motion compensation prediction is performed on blocks included in the merge processing area, motion information candidates derived from the blocks can be added to the motion information table in a predefined order. Here, a predefined definition The ordered order may be determined according to the scanning order of the coding blocks in the merge processing area or the coding tree unit. The scanning order may be at least one of raster scan, horizontal scan, vertical scan, or zigzag scan. Or, a predefined order Is the movement information of each block or the same movement 2020/175915 1»（：1^1{2020/002754 It can be determined based on the number of blocks with information.

[301] Or, a motion information candidate including one-way motion information, and two-way motion information.

It is possible to add motion information candidates including two-way motion information to the motion information table earlier than the included motion information candidates. Conversely, motion information candidates including bi-directional motion information may be added to the motion information table earlier than motion information candidates containing unidirectional motion information.

[302] Alternatively, motion information candidates can be added to the motion information table in the order of high frequency of use or low frequency of use in the merge processing area or coding tree unit.

[303] If the current block is included in the merge processing area and the number of merge candidates included in the merge candidate list of the current block is less than the maximum number, the motion information candidate included in the motion information table can be added to the merge candidate list. At this time, motion information candidates derived from blocks included in the same merge processing area as the current block may be set not to be added to the merge candidate list of the current block.

[304] Alternatively, if the current block is included in the merge processing area, the motion information candidates included in the motion information table can be set not to be used. That is, the number of merge candidates included in the merge candidate list of the current block is maximum. Even if it is smaller than the number, motion information candidates included in the motion information table may not be added to the merge candidate list.

[305] As another example, a motion information table for a merge processing area or a coding tree unit can be constructed. This motion information table serves to temporarily store motion information of blocks included in the merge processing area. General motion In order to distinguish between the information table and the motion information table for the merge processing area or the coding tree unit, the motion information table for the merge processing area or the coding tree unit will be referred to as a temporary motion information table. The motion information candidate will be referred to as a temporary motion information candidate.

25 is a diagram showing a temporary motion information table.

[307] A temporary motion information table for the coding tree unit or the merge processing area can be configured. When motion compensation prediction is performed for the current block included in the coding tree unit or the merge processing area, the motion information of the block is Information table

추가하지 않을수있다.대신,상기블록으로부터유도된임시 모션정보후보를임시모션정보테이블 HmvpMERCandList에추가할수있다. 즉,임시모션정보테이블에추가된임시모션정보후보는모션정보테이블에 추가되지 않을수있다.이에따라,모션정보테이블은현재블록을포함하는 코딩트리유닛또는머지 처리 영역에포함된블록들의모션정보를기초로 유도된모션정보후보를포함하지 않을수있다.

It may not be added. Instead, the temporary motion information candidate derived from the block can be added to the temporary motion information table HmvpMERCandList. That is, the temporary motion information candidate added to the temporary motion information table may not be added to the motion information table. Accordingly, the motion information table is based on motion information of blocks included in the coding tree unit including the current block or the merge processing area. It may not include motion information candidates induced by.

[308] Alternatively, it is possible to add only the motion information of some of the blocks included in the head processing area to the temporary motion information table. For example, only blocks in the defined position in the merge processing area are used to update the motion information table. The predefined position is the block located in the upper left corner of the merge processing area, and in the upper right corner. 2020/175915 1»（：1^1{2020/002754 Block located, Block located at the bottom left, Block located at the bottom right, Block located at the center, Block adjacent to the right boundary, or It may contain at least one of the blocks. For example, only the motion information of the block adjacent to the lower right corner in the merge processing area may be added to the temporary motion information table, and the motion information of other blocks may not be added to the temporary motion information table. .

[309] The maximum number of temporary motion information candidates that the temporary motion information table can contain may be set equal to the maximum number of motion information candidates that the motion information table can contain. Or, the temporary motion information candidate that the temporary motion information table can contain. The maximum number of can be determined according to the size of the coding tree unit or merge processing area, or the maximum number of temporary motion information candidates that the temporary motion information table can contain is less than the maximum number of motion information candidates that the motion information table can contain. Can be set.

[310] The current block included in the coding tree unit or the merge processing area can be set not to use the temporary motion information table for the corresponding coding tree unit or the merge processing area. That is, included in the merge candidate list of the current block. If the number of merge candidates is less than the threshold value, the motion information candidates included in the motion information table may be added to the merge candidate list, and the temporary motion information candidates included in the temporary motion information table may not be added to the merge candidate list. Accordingly, motion information of the same coding tree unit as the current block or of other blocks included in the same merge processing area may not be used for the motion compensation prediction of the current block.

[311] When encoding/decoding of all blocks included in the coding tree unit or merge processing area is completed, the motion information table and the temporary motion information table can be merged.

26 is a diagram showing an example of merging the motion information table and the temporary motion information table.

[313] When coding/decoding of all blocks included in the coding tree unit or merge processing area is completed, the temporary motion information candidate included in the temporary motion information table is updated to the motion information table as in the example shown in FIG. can do.

[314] At this time, the temporary motion information candidates included in the temporary motion information table may be added to the motion information table in the order they are inserted into the temporary motion information table (ie, ascending or descending index value).

[315] As another example, according to a predefined order, the temporary motion information candidates included in the temporary motion information table can be added to the motion information table. Here, the predefined order is the merge processing area or the coding within the coding tree unit. It can be determined according to the scanning order of blocks. The scanning order may be at least one of raster scan, horizontal scan, vertical scan, or zigzag scan. Or, the predefined order is the motion information of each block or having the same motion information. It can be determined based on the number of blocks.

[316] Alternatively, the temporary motion information candidate including the one-way motion information may be added to the motion information table earlier than the temporary motion information candidate including the bi-directional motion information. 2020/175915 1»（：1^1{2020/002754 Conversely, it is also possible to add a temporary motion information candidate containing bidirectional motion information to the motion information table before the temporary motion information candidate containing one-way motion information.

Alternatively, temporary motion information candidates may be added to the motion information table in the order of high frequency of use or low frequency of use in the merge processing area or the coding tree unit.

[318] In the case of adding the temporary motion information candidates included in the temporary motion information table to the motion information table, a redundancy test for the temporary motion information candidates can be performed. For example, the temporary motion information included in the temporary motion information table can be performed. If the same motion information candidate as the information candidate is previously stored in the motion information table, the temporary motion information candidate may not be added to the motion information table. At this time, the redundancy check is performed on some of the motion information candidates included in the motion information table. For example, a redundancy test may be performed on motion information candidates whose index is above or below the threshold. For example, a motion information candidate with an index above a predefined value and In the same case, the temporary motion information candidate may not be added to the motion information table.

[319] In the same coding tree unit or in the same merge processing area as the current block

The motion information candidate derived from the block can be restricted from being used as the merge candidate of the current block. To this end, the address information of the block can be additionally stored for the motion information candidate. The address information of the block is, the location of the block, the block's address information. Address, block index, position of merge processing area containing block, address of merge processing area containing block, index of merge processing area containing block, position of coding tree area containing block, block included It may contain at least one of the address of the coded tree area or the index of the coded tree area containing the block.

[32 this

[321] The coding block is divided into a plurality of prediction units, and each of the divided prediction units is

It is possible to make predictions, where the prediction unit represents the basic unit for performing the prediction.

[322] A coding block may be divided using at least one of a vertical line, a horizontal line, a diagonal line, or a diagonal line. The prediction units divided by the division line may have a shape such as a triangle, a square, a trapezoid, or a pentagon. For example, coding The block may be divided into two triangular-shaped prediction units, two trapezoidal-shaped prediction units, two square-shaped prediction units, or one triangular-shaped prediction unit and one pentagonal-shaped prediction unit.

[323] Information for determining at least one of the number, angle, or position of the lines dividing the coding block may be signaled through the bitstream. For example, information indicating any one of the partition type candidates of the coding block is a bit bit. One of the multiple line candidates signaled through a stream or splitting a coding block 2020/175915 1»（：1^1{2020/002754 Specific information can be signaled through a bitstream. For example, index information indicating any one of a plurality of line candidates can be signaled through the bitstream. have.

[324] Each of the multiple line candidates may be virtually different from at least one of the angles or positions. The number of line candidates for which the current block is available may be determined based on the size and shape of the current block, the number of available merge candidates, or whether a neighboring block at a specific location can be used as a merge candidate.

[325] Alternatively, information for determining the number or type of line candidates

For example, a 1-bit flag can be used to determine whether a diagonal line with an angle larger than the diagonal and/or a diagonal line with an angle smaller than the diagonal line can be used as a line candidate. It can be signaled at the picture or sequence level.

[326] Alternatively, based on at least one of the intra prediction mode, the inter prediction mode, the position of the available merge candidates, or the segmentation pattern of the neighboring block, at least one of the number, angle, or position of the lines dividing the coding block It can be determined adaptively.

[327] When a coding block is divided into a plurality of prediction units, intra prediction or inter prediction can be performed on each of the divided prediction units.

27 is a diagram showing an example of dividing a coding block into a plurality of prediction units by using a diagonal line.

[329] As in the 如 and (non-shown example in Fig. 27), the coding block is

It can be divided into two triangular shape prediction units.

[33] In Fig. 27, the diagonal line connecting the two vertices of the coding block is used.

It is shown that the coding block is divided into two prediction units. However, the coding block can be divided into two prediction units by using a diagonal line in which at least one end of the line does not pass the vertex of the coding block.

28 is a diagram illustrating an example of dividing a coding block into two prediction units.

[332] As in the example shown in 如 and） of Fig. 28, the coding block can be divided into two prediction units by using a diagonal line where both ends are in contact with the upper and lower boundaries of the coding block.

[333] Or, as in the example shown in Figure 28 (0) and ), both ends of the coding block

The coding block can be divided into two prediction units by using the diagonal lines tangent to the left and right borders.

[334] Alternatively, the coding block can be divided into two prediction units of different sizes. For example, by setting the diagonal line dividing the coding block to be in contact with two boundaries forming one vertex, the coding block can be divided into two different sized prediction units. It can be divided into prediction units of.

29 shows examples of dividing a coding block into a plurality of prediction units having different sizes. 2020/175915 1»（：1^1{2020/002754

[336] As in the example of FIG. 29 and (as in the non-shown example, the upper left and lower right of the coding block

By setting the connecting diagonal line to pass through the left, right, upper, or lower borders instead of passing the upper left corner or lower right corner of the coding block, the coding block can be divided into two prediction units of different sizes.

[337] Or, as in the example shown in (0) and) of FIG. 29, the upper right and

By setting the diagonal line connecting the lower left corner to pass through the upper left corner or the lower right corner of the coding block, instead of passing through the left boundary, right boundary, upper boundary or lower boundary, the coding block is divided into two prediction units with different sizes. can do.

예측유닛’이라 [338] Each of the prediction units generated by dividing the coding block

'Prediction Unit'

As an example, in the examples shown in Figs. 27 to 29, may be defined as a first prediction unit and 1 2 as a second prediction unit. The first prediction unit is a sample located at the lower left of the coding block. Or the sample located in the upper left corner

It refers to a prediction unit including, and the second prediction unit may mean a prediction unit including a sample located at the upper right or a sample located at the lower right of the coding block.

[339] Contrary to the above, a prediction unit including a sample located at the top right or a sample located at the bottom right is defined as the first prediction unit, and includes a sample located at the bottom left or a sample located at the top left in the coding block. The predicting unit can be defined as a second example, namely unit.

[34] The division of the coding block using horizontal lines, vertical lines, diagonal lines, or diagonal lines can be referred to as prediction unit partitioning. The prediction units generated by applying the prediction unit partitioning, according to their shape, are triangular prediction units and square prediction units. It can be referred to as a unit or pentagonal prediction unit.

[341] In the embodiments described below, it is assumed that the coding block is divided using a diagonal line. In particular, dividing the coding block into two prediction units using a diagonal line will be referred to as diagonal partitioning or triangular partitioning. However, even when the coding block is divided using a vertical line, a horizontal line, or a diagonal line of an angle different from the diagonal line. , The prediction units can be encoded/decoded according to the embodiments described below. That is, matters related to encoding/decoding of the triangular prediction unit to be described later can also be applied to the encoding/decoding of the square prediction unit or the pentagonal prediction unit.

[342] Whether to apply prediction unit partitioning to a coding block, the slice type, the maximum number of merge candidates that the merge candidate list can contain, the size of the coding block, the shape of the coding block, the prediction coding mode or the parent node of the coding block It can be determined on the basis of at least one of the aspects of the division.

[343] For example, it is possible to determine whether to apply prediction unit partitioning to the coding block based on whether the current slice is: 8 types. Prediction unit partitioning can be allowed only when the current slice is: 8 types. .

[344] Or, whether the maximum number of merged candidates included in the merger candidate list is 2 or more. 2020/175915 1»（：1^1{2020/002754 You can decide whether to apply prediction unit partitioning to the coding block or not. The prediction unit partitioning is the maximum number of merge candidates included in the merge candidate list. It can be allowed only if there are more than two.

[345] Alternatively, if at least one of the width or height of the hardware implementation is larger than 64, there is a disadvantage that the data processing unit of size 64 4 has redundant access. Accordingly, at least one of the width or height of the coding block is the threshold value. If it is larger than, it may not be allowed to divide the coding block into a plurality of prediction units. For example, if at least one of the height or width of the coding block is greater than 64 (e.g., at least one of the width or height is 128) Right), prediction unit partitioning may not be used.

[346] Alternatively, considering the maximum number of samples that can be processed at the same time according to the hardware implementation, prediction unit partitioning may not be allowed for a coding block with a number of samples greater than a threshold. For example, a coding tree block with a number of samples greater than 4096 may not be allowed. May not allow prediction unit partitioning.

[347] Alternatively, prediction unit partitioning may not be allowed for a coding block in which the number of samples included in the coding block is smaller than the threshold. For example, if the number of samples included in the coding block is less than 64, the coding block It can be set so that prediction unit partitioning is not applied.

비율로서 결정될수있다. [348] Or, whether to apply prediction unit partitioning to the coding block based on whether the width and height ratio of the coding block is less than the first threshold or whether the width and height ratio of the coding block is greater than the second threshold. Here, the width and height ratio of the coding block \vhRatio is the width of the coding block as shown in Equation 2 below.

It can be determined as a ratio.

[349] [Equation 2]

Well 1 yo幻 0 =（That 1? /<=1?11

[35] The second threshold may be an reciprocal of the first threshold. For example, when the first threshold is no, the second threshold may be 1.

[351] The width and height ratio of the coding block exists between the first threshold and the second threshold.

Prediction unit partitioning can be applied to the coding block only in case.

[352] Alternatively, prediction unit partitioning can be used only when the width and height ratio of the coding block is smaller than the first threshold or larger than the second threshold. For example, when the first threshold is 16, 64x4 or 4x64 Prediction unit partitioning may not be allowed for sized coding blocks.

[353] Alternatively, it is possible to determine whether to allow the prediction unit partitioning based on the division pattern of the parent node. For example, when the parent node encoding block is divided based on quadtree division, the leaf node encoding block includes the prediction unit. Partitioning can be applied. On the other hand, if the parent node encoding block is divided on the basis of binary retrieval or triple tree division, the prediction unit partitioning is not allowed in the leaf node encoding block. 2020/175915 1»（：1^1{2020/002754 can be set.

Alternatively, based on the prediction coding mode of the coding block, it is possible to determine whether to allow prediction unit partitioning. For example, when a coding block is coded by intra prediction, when a coding block is coded by inter prediction, or coding Prediction unit partitioning is allowed only when the block is coded with a defined inter prediction mode. Here, the predefined intere-side mode may include at least one of a merge mode, a motion vector prediction mode, an affine merge mode, and an afine motion vector prediction mode.

[355] Alternatively, based on the size of the parallel processing area, it is possible to determine whether to allow prediction unit partitioning. For example, if the size of the coding block is larger than the size of the parallel processing area, the prediction unit partitioning may not be used. .

[356] Considering two or more of the conditions listed above, it is also possible to decide whether to apply prediction unit partitioning to the coding block.

[357] As another example, information indicating whether to apply prediction unit partitioning to a coding block may be signaled through a bitstream. The information may be signaled at the sequence, picture, slice, or block level. For example, A flag triangle_paGtition_flag indicating whether prediction unit partitioning is applied to the coding block can be signaled at the coding block level.

[358] When it is decided to apply prediction unit partitioning to a coding block, the coding block is

Information indicating the number of lines to be divided or the position of the lines may be signaled through the bitstream.

[359] As an example, when a coding block is divided by a diagonal line, information indicating the direction of a diagonal line for dividing the coding block may be signaled through the bitstream. For example, a triangle_paGtition_type_flag indicating the direction of the diagonal line may be signaled.

It can be signaled through a bitstream. The flag indicates whether the coding block is divided by a diagonal line connecting the top left and bottom right, or whether it is divided by a diagonal line connecting the top left and bottom left. It is referred to as a left triangular partition type that divides the coding block by, and the coding block is divided by a diagonal line connecting the upper right and the lower left as a right triangular partition type. For example, if the value of the flag is 0, the value of the flag is 0. It indicates that the partition type of the coding block is the left triangular partition type, and the flag value of 1 indicates that the partition type of the coding block is the right triangular partition type.

[36 additionally, information or coding indicating whether the size of the prediction units is the same

Information indicating the position of the diagonal line dividing the block can be signaled through the bitstream. For example, if the information indicating the size of the prediction units indicates that the size of the prediction units is the same, the information indicating the position of the diagonal line is encoded. Is omitted, and the coding block can be divided into two prediction units using a diagonal line passing through the two vertices of the coding block, whereas information indicating the size of the prediction units indicates that the size of the prediction units is not the same, 2020/175915 1»（：1^1{2020/002754 Based on the information indicating the position of the diagonal line, the position of the diagonal line that divides the coding block can be determined. For example, the left triangular partition type is applied to the coding block. In this case, the location information may indicate whether the diagonal line is in contact with the left boundary and the lower boundary of the coding block, or whether the upper boundary and the right boundary are in contact. Alternatively, when the right triangular partition type is applied to the coding block, the location information may indicate whether the diagonal line is in contact with the right boundary and the lower boundary of the coding block, or whether the upper boundary and the left boundary are in contact.

[361] Information indicating the partition type of the coding block can be signaled at the coding block level. Accordingly, the partition type can be determined for each coding block to which prediction unit partitioning is applied.

[362] As another example, information indicating the partition type for a sequence, picture, slice, tile, or coding tree unit may be signaled. In this case, a sequence, picture, slice, tile, or coding in which diagonal partitioning within a coding tree unit is applied The partition types of blocks can be set identically.

[363] Alternatively, the information for determining the partition type is encoded and signaled for the first coding unit to which the prediction unit partitioning in the coding tree unit is applied, and the second and subsequent coding units to which the prediction unit partitioning is applied are the first coding unit and You can configure it to use the same partition type.

[364] As another example, based on the partition type of the neighboring block, the partition type of the coding block

Here, the neighboring block is a neighboring block adjacent to the upper left corner of the coding block, a neighboring block adjacent to the upper right corner, a neighboring block adjacent to the lower left corner, a neighboring block located at the top, or a neighboring block located at the left. It may contain at least one of the neighboring blocks. For example, the partition type of the current block may be set the same as the partition type of the neighboring block. Or, whether the left triangular partition type is applied to the upper left neighboring block, the upper right neighboring block, or The partition type of the current block can be determined based on whether the right triangular partition type is applied to the lower left neighboring block.

[365] In order to perform the motion prediction compensation for the first prediction unit and the second prediction unit, it is possible to induce motion information of each of the first prediction unit and the second prediction unit. At this time, the first prediction unit and the second prediction unit. The movement information of the unit can be derived from the merge candidates included in the merge candidate list. In order to distinguish between the general merge candidate list and the merge candidate list used to derive the movement information of the prediction units, the movement information of the prediction units is derived. The merge candidate list will be referred to as a split mode merge candidate list or a triangular merge candidate list. In addition, the merge candidate included in the split mode merge candidate list will be referred to as a split mode merge candidate or a triangle merge candidate. It is also included in the idea of the present invention that the candidate induction method and the merge candidate list construction method are used in the split mode merge candidate and the split mode merge candidate list construction method. 2020/175915 1»（：1^1{2020/002754

[366] Information for determining the number of maximum split mode merge candidates that the split mode merge candidate list can contain can be signaled through the bitstream. The above information includes the maximum number of merge candidates that the merge candidate list can contain and the split mode candidate list. You can indicate the difference between the maximum number of split mode merge candidates that the merge candidate list can contain.

[367] Spatial neighboring blocks and temporal neighbors of the coding block as the partition mode merge candidate

Can be derived from blocks.

FIG. 30 is a diagram showing neighboring blocks used to induce a split mode merge candidate.

[369] Split mode merge candidates are included in the neighboring block located at the top of the coding block, the neighboring block located at the left of the coding block, or a picture different from the coding block.

It can be derived by using at least one of the collocated blocks. The upper neighboring block is a block containing samples 江（¾+（¾\¥-1, （¾-1）） located at the top of the coding block, It may include at least one of a block containing the sample 江 (¾+ example (¾-1) at the top, or a block containing the sample 江 (¾-1, （¾-1)) located at the top of the coding block. The left neighboring block is a block containing sample 江 (¾-1, （¾+幻^-1) located on the left side of the coding block or sample 江幻5-1, （¾+幻） located on the left side of the coding block. It can contain at least one of the blocks containing 出). The collocated block is a block containing a sample 江幻5+example 幻5+0出) adjacent to the upper right corner of the coding block in the collocated picture, or a sample 江 (¾/2, located in the center of the coding block). It can be determined by any one of the blocks containing (¾/2).

[37] Searches for neighboring blocks in a predefined order, and according to the predefined order, it is possible to construct a split mode merge candidate list from a split mode merge candidate. For example: 81, company, yes, show 0, 00 You can configure the list of split mode merge candidates by searching for split mode merge candidates in the order of ,: 82 and 0.

[371] The motion information of the prediction units is based on the split mode merge candidate list.

It can be derived, i.e., prediction units can share a single split mode merge candidate list.

[372] In order to guide the motion information of the prediction unit, it is added to the split mode merge candidate list.

Information for specifying at least one of the included split mode merge candidates may be signaled through the bitstream. For example, at least one of the split mode merge candidates may be signaled.

Can be signaled through.

[373] The index information is based on the merge candidate of the first prediction unit and the merge candidate of the second prediction unit.

The combination can be specified. As an example, the following table 3 is an example showing the combination of the merged candidates according to index information 111 6_1;] 1¾16_:1 (1). 2020/175915 1»（：1^1{2020/002754

[374] [Table 3]

[375] Index information 111 when 0] urine 1¾1 （If the urine value is 1, the movement of the first prediction unit

The information indicates that the index is derived from the merge candidate in 1, and the motion information of the second prediction unit is derived from the merge candidate in 0. Index information

정보를유도하기위한분할모드머지 후보가결정될수있다.인덱스정보에의해,대각파티셔닝이적용되는코딩 블록의파티션타입을결정할수도있다.즉,인덱스정보는,제 1예측유닛의 머지후보,제 2예측유닛의머지후보및코딩블록의분할방향의조합을 특정할수있다.인덱스정보에의해코딩블록의파티션타입이결정되는경우, 코딩블록을분할하는대각선의방향을나타내는정보 Split to induce motion information

The partitioning mode merge candidate for inducing the information can be determined. According to the index information, the partition type of the coding block to which the diagonal partitioning is applied can be determined. That is, the index information is the merge candidate of the first prediction unit, the second It is possible to specify a combination of the prediction unit's merge candidate and the direction of division of the coding block When the partition type of the coding block is determined by the index information, information indicating the direction of the diagonal to divide the coding block

The triangle_paGtition_type_flag may not be coded. Table 4 shows the index information 1116 6_1::1' 1¾16_:1 (partition type of the urine encoding block. 2020/175915 1»（：1^1{2020/002754

[376] [Table 4]

[377] Variable 1¾ min _£ 1) Urinary value of 0 indicates that the left triangular partition type is applied to the coding block.

The variable 1¾ min _£ 1) Urine 1 indicates that the right triangular partition type is applied to the coding block. Table 3 and Table 4 are combined, index information 111 6_1;] Urine 1¾16_:1 (1 is the 1st prediction unit's The combination of the merge candidate, the merge candidate of the second prediction unit, and the direction of division of the coding block can be specified. As another example, only the index information for either the first prediction unit and the second prediction unit is signaled, and Based on the above index information, it is possible to determine the index of the merge candidate for the other one of the first prediction unit and the second prediction unit. For example, index information 1116 6_1::1 indicating the index of any one of the split mode merge candidates. '1¾16_: 1 (Based on urine, it is possible to determine the merge candidate of the first example, that is, the unit's merger. And, the above 1116 _1;] urine 1¾16_:1 (Based on 1, the second prediction unit's merge candidate can be specified. As an example, the merge candidate of the second prediction unit can be derived by adding or subtracting an offset to the index information 111 = 6_1;] urine 1¾16_: 1 (1). The offset may be an integer such as 1 or 2. ,The second prediction unit's merge candidate is 111 6_1; 1¾16_:1 (can be determined as a split mode merge candidate having the value of 1 plus 1 as an index. If 111 6_1;] 1¾16_:1 (1 temporary split mode Among merge candidates, if the index value points to the split mode merge candidate with the largest index, the split mode merge candidate with an index of 0 or 1116 6_1;:1' 1¾16_:1 (from the split mode merge candidate with the value of 1 in urine as an index The movement information of the second prediction unit can be derived.

[378] Or, with the split mode merge candidate of the first prediction unit specified by the index information

The motion information of the second prediction unit can be derived from the split mode merge candidate having the same reference picture. Here, the split mode merge candidate having the same reference picture as the split mode merge candidate of the first prediction unit At least one of the split mode merge candidate and the 1 _^ 0 reference picture or the 1 _^ 1 reference picture may represent the same split mode merge candidate. The split mode merge candidate and the reference picture of the first prediction unit are 2020/175915 1»（：1^1{2020/002754 When there are multiple candidates for the same division mode merge, at least one of whether the merge candidate contains bidirectional motion information, or the difference between the index information of the merge candidate and the index information. You can choose any one based on.

[379] As another example, the index information for each of the first prediction unit and the second prediction unit

기초로 결정되는분할모드머지후보로부터유도되고,제 2예측유닛의움직임 정보는 제 2인덱스정보 2!1（1_111_{£ £}_:1（뇨를기초로결정되는분할모드머지후보로부터 유도될수있다. As an example, when the first index information 1 _111 for determining the split mode merge candidate of the first prediction unit (Urine and second index information 211 for determining the split mode merge candidate of the second prediction unit) 1_1 _£ _:1 (Urine can be signaled through the bitstream. The movement information of the first prediction unit is the first index.

It is derived from the split mode merge candidate determined on the basis, and the movement information of the second prediction unit can be derived from the split mode merge candidate determined on the basis of the second index information 2!1 (1_111 _{£ £} _:1 (urea).

[38 is the first index information 1 _1116¾6_: 1 (Urine can indicate any one index among the split mode merge candidates included in the split mode merge candidate list. The split mode merge candidate of the first prediction unit is

May be decided as a merger candidate.

후보는제 2예측 유닛의분할모드머지후보로이용가능하지 않도록설정될수있다.이에 따라, 제 2예측유닛의 제 2인덱스정보 211（1_1116 6_:1（뇨는제 1인덱스정보가가리키는 분할모드머지후보를제외한잔여분할모드머지후보들중어느하나의 인덱스를나타낼수있다.제 2인덱스정보 211（1_1116 6_:1（뇨의 값이제 1인덱스 값보다작은경우,제 2예측유닛의분할모드머지후보는

1116 6_:1（뇨가나타내는인덱스정보를갖는분할모드머지 후보로결정될수있다.반면,제 2인덱스정보 211（1_111때 （뇨의 값이제 1인덱스 정보 1 _111_£ （뇨의값과같거나큰경우,제 2예측유닛의분할모드머지 후보는제 2인덱스정보 2!1（1_111_{£ £}_:1（뇨의 값에 1을더한값을인덱스로갖는 분할모드머지후보로결정될수있다. [381] Index 1

The candidate may be set not to be available as a split mode merge candidate of the second prediction unit. Accordingly, the second index information 211 (1_1116 6_:1) of the second prediction unit (regarding the division mode merge indicated by the first index information) Any one of the remaining split mode merge candidates excluding the candidate can be displayed. Second index information 211 (1_1116 6_:1 (if the urine value is less than the first index value, the second prediction unit's split mode merge candidate)

1116 6_:1 (Can be determined as a partition mode merge candidate with the index information indicating urine. On the other hand, the second index information 211 (when 1_111) (the urine value is now 1 index information 1 _111 _£ （equal to or greater than the urine value In this case, the division mode merge candidate of the second prediction unit may be determined as a division mode merge candidate having the second index information 2!1 (1_111 _{£ £} _:1 (a urine value plus 1) as an index.

[382] Alternatively, it is possible to determine whether or not to signal the second index information according to the number of split mode merge candidates included in the split mode merge candidate list. For example, the maximum number of split mode merge candidates that the split mode merge candidate list can include. If the number does not exceed 2, the signaling of the second index information may be omitted. If the signaling of the second index information is omitted, an offset is added or subtracted to the first index information to be eligible for the second division mode merge. As an example, if the maximum number of split mode merge candidates that the split mode merge candidate list can include is 2, and the first index information points to index 0, 1 is added to the first index information, and the second Alternatively, if the maximum number of split mode merge candidates that the split mode merge candidate list can include is two, and the first index information points to 1, the first index information is divided by 1 Two-split mode merge candidates can be induced.

[383] Or, if the signaling of the second index information is omitted, the second index information is defaulted. 2020/175915 1»（：1^1{2020/002754 value can be set. Here, the default value can be 0. The first index information and the second index information are compared, and the second division mode merge candidate is induced. For example, if the second index information is smaller than the first index information, the index 0 index merge candidate is set as the second split mode merge candidate, and the second index information is equal to or greater than the first index information, The index 1 merge candidate can be set as the second split mode merge candidate.

[384] If the split mode merge candidate has one-way motion information, the one-way motion information of the split-mode merge candidate can be set as the motion information of the prediction unit. On the other hand, the split-mode merge candidate has two-way motion information.

후보의 인덱스또는타예측유닛의움직임 정보를기초로결정될수있다. Only one of the motion information can be set as the motion information of the prediction unit. It is determined whether the motion information or the motion information is to be taken.

It can be determined based on the candidate index or motion information of other prediction units.

움직임 정보를예측유닛의 니움직임 정보로설정할수있다.반면,분할모드머지후보의 인덱스가홀수인 경우,예측유닛의 1움직임 정보를 0으로설정하고,분할모드머지후보의

움직임 정보를 0으로설정할수있다.위와반대로,분할모드머지후보의 인덱스가짝수인경우,분할모드머지후보의 ）움직임 정보를예측유닛의 ） 움직임 정보로설정하고,분할모드머지후보의 인덱스가홀수인경우,분할 모드머지후보의니움직임정보를예측유닛의니움직임정보로설정할수도 있다.또는,제 1예측유닛에 대해서는분할모드머지후보가짝수인경우,분할 모드머지후보의

정보를제 1예측유닛의 ）움직임 정보로설정하는 반면,제 2예측유닛에 대해서는분할모드머지후보가홀수인경우,분할모드 머지후보의

정보를제 2예측유닛의니움직임정보로설정할수 있다. [385] For example, if the index of the split mode merge candidate is even number, the movement information of the prediction unit is set to 0, and the split mode merge candidate's index is set to 0.

The motion information can be set as the motion information of the prediction unit. On the other hand, if the index of the split mode merge candidate is odd, the one motion information of the prediction unit is set to 0, and the split mode merge candidate's index is 0.

The motion information can be set to 0. Contrary to the above, if the index of the split mode merge candidate is even, the motion information of the split mode merge candidate is set as the motion information of the prediction unit, and the index of the split mode merge candidate is odd. In the case of, it is also possible to set the tooth motion information of the split mode merge candidate as the tooth motion information of the prediction unit. Or, if the number of split mode merge candidates is even for the first prediction unit, the split mode merge candidate

While the information is set as the movement information of the first prediction unit, for the second prediction unit, if the number of split mode merge candidates is odd, the split mode merge candidate's

The information can be set as the motion information of the second prediction unit.

[386] Alternatively, if the first prediction unit has movement information, the movement information of the second prediction unit can be set to 0, and the motion information of the split mode merge candidate can be set as the knee information of the second prediction unit. On the other hand, when the first prediction unit has the movement information, the movement information of the second prediction unit can be set to 0, and the movement information of the split mode merge candidate can be set as the movement information of the second prediction unit.

[387] The split mode merge candidate list for inducing the motion information of the first prediction unit and the split mode merge candidate list for inducing the motion information of the second prediction unit may be set differently.

[388] As an example, split mode merge candidate based on the index information for the first prediction unit

When a split mode merge candidate for inducing the motion information of the first prediction unit in the list is specified, the motion information of the second prediction unit is split mode merge, including the remaining split mode merge candidates excluding the split mode merge candidate indicated by the index information. It can be derived using a list. Specifically, the movement information of the second prediction unit can be derived from any one of the remaining division mode merge candidates. 2020/175915 1»（：1^1{2020/002754

[389] Accordingly, the number of maximum split mode merge candidates included in the split mode merge candidate list of the first prediction unit and the maximum split mode merge candidate list included in the split mode merge candidate list of the second prediction unit may be different. For example, if the split mode merge candidate list of the first prediction unit contains M merge candidates, the split mode merge candidate list of the second prediction unit excludes the split mode merge candidate indicated by the index information of the first prediction unit. ^Can contain 1-1 merge candidates.

[39] As another example, the merge candidate of each prediction unit is guided based on the neighboring blocks adjacent to the coding block, but the availability of the neighboring block can be determined by considering the shape or location of the prediction unit.

31 is a diagram for explaining an example of determining availability of neighboring blocks for each prediction unit.

[392] A neighboring block not adjacent to the first prediction unit may be set as unavailable for the first prediction unit, and a neighboring block not adjacent to the second prediction unit may be set as unavailable for the second prediction unit.

[393] As an example, as in the example shown in Fig. 31), the left triangular partition in the coding block

When the type is applied, the blocks show 1, show 0 and show 2 adjacent to the first prediction unit among the neighboring blocks adjacent to the coding block are available for the first prediction unit, whereas block: 80 and the second are unavailable for the first prediction unit. Accordingly, the split mode merge candidate list for the first prediction unit includes the split mode merge candidates derived from block show 1, show 0 and show 2, while block 60 and the split mode merge derived therefrom. Candidates may not be included.

[394] [394] In the case where the left triangular partition type is applied to the coding block as in the non-shown example of Fig. 31, the block example adjacent to the second prediction unit and the second prediction unit can be used, whereas the block show 1 , Show 0 and Show 2 may be determined to be unavailable for the second prediction unit. Accordingly, the split mode merge candidate list for the second prediction unit includes block examples and split mode merge candidates derived therefrom, while the block show Split mode merge candidates derived from 1, Show 0 and Show 2 may not be included.

[395] Accordingly, the number of split mode merge candidates that the prediction unit can use or the range of split mode merge candidates can be determined based on at least one of the location of the prediction unit or the partition type of the coding block.

[396] As another example, only one of the first prediction unit and the second prediction unit can be used in the merge mode.

In addition, the motion information of the other one of the first and second prediction units is set to be the same as the motion information of the prediction unit to which the merge mode is applied, or the motion information of the prediction unit to which the merge mode is applied is refined. Can be induced by

[397] As an example, based on the split mode merge candidate, the motion vector of the first prediction unit and the reference picture index can be derived, and the motion vector of the second prediction unit can be derived by refining the motion vector of the first prediction unit. For example, the motion vector of the second prediction unit is 2020/175915 1»（：1^1{2020/002754 It can be derived by adding or subtracting the refine motion vector {Rx, Ry} to the motion vector {mvDlLXx, mvDlLXy} of the first prediction unit. Reference to the second prediction unit. The picture index may be set equal to the reference picture index of the first prediction unit.

[398] Information for determining a refine motion vector indicating a difference between the motion vector of the first prediction unit and the motion vector of the second prediction unit may be signaled through the bitstream. The above information indicates the size of the refine motion vector. It may include at least one of the indicated information or information indicating the sign of the refine motion vector.

[399] Or, one of the location of the prediction unit, the index, or the partition type applied to the coding block.

Based on at least one, it is possible to derive the sign of the refine motion vector.

[40 is another example, one of the first and second prediction units can signal a motion vector and a reference picture index. The other one of the first and second prediction units is the signal above. It can be derived by refining the ringed motion vector.

[401] For example, based on the information signaled from the bitstream, the motion vector of the first prediction unit and the reference picture index can be determined. The motion vector of the first prediction unit is refined to The motion vector may be derived. For example, the motion vector of the second prediction unit may be derived by adding or subtracting the refine motion vector {Rx, Ry} to the motion vector {mvDILXx, mvDILXy} of the first prediction unit. The reference picture index of the second prediction unit may be set equal to the reference picture index of the first prediction unit.

[402] As another example, only one of the first predicting unit and the second predicting unit is the merge mode.

In addition, the motion information of the other one of the first prediction unit and the second prediction unit may be derived based on the motion information of the prediction unit to which the merge mode is applied. For example, the motion vector of the first prediction unit The symmetric motion vector may be set as the motion vector of the second prediction unit, where the symmetric motion vector is the same size as the motion vector of the first prediction unit, but at least one of the X-axis or y-axis components is opposite. The in-motion vector or the scaled vector obtained by scaling the motion vector of the first prediction unit may have the same size, but at least one of the X-axis component or the y-axis component may mean a motion vector whose sign is opposite. When the motion vector of the first prediction unit is (MVx, MVy), the motion vector of the second prediction unit is (MVx, -MVy), (-MVx, MVy) or (-MVx,-which is the symmetrical motion vector of the motion vector. MVy) can be set.

[403] Among the first and second prediction units, the reference picture index of the prediction unit to which the merge mode is not applied may be set equal to the reference picture index of the prediction unit to which the merge mode is applied. Or, the merge mode is not applied. The reference picture index of the non-prediction unit may be set to a predefined value. Here, the predefined value may be the smallest index or the largest index in the reference picture list. Or, the merge 2020/175915 1»（：1^1{2020/002754 Information specifying the reference picture index of the prediction unit to which the mode is not applied can be signaled through the bitstream. Or, the reference picture of the prediction unit to which the merge mode is applied. The reference picture of the prediction unit to which the merge mode is not applied can be selected from the reference picture list different from the accelerated reference picture list. For example, when the reference picture of the prediction unit to which the merge mode is applied is selected from the L0 reference picture list, the merge mode The reference picture of the prediction unit to which the merge mode is not applied can be selected from the L1 reference picture list. At this time, the reference picture of the prediction unit to which the merge mode is not applied is the output order between the reference picture of the prediction unit to which the merge mode is applied and the current picture ( Picture Order Count, POC) can be derived based on the difference. For example, when the reference picture of the prediction unit to which the merge mode is applied is selected from the L0 reference picture list, the difference value from the current picture in the L1 reference picture list is applied to the merge mode. A reference picture equal to or similar to a difference value between the reference picture of the prediction unit and the current picture can be selected as the reference picture of the prediction unit to which the merge mode is not applied.

[404] When the size of the output order difference value between the reference picture of the first prediction unit and the current picture and the output order difference value between the reference picture and the current picture of the second prediction unit are different, scaling of the prediction unit to which the merge mode is applied The symmetrical motion vector of the old motion vector can be set as the motion vector of the prediction unit to which the merge mode has not been applied. At this time, scaling can be performed based on the output order difference value between each reference picture and the current picture.

As another example, after inducing a motion vector of each of the first prediction unit and the second prediction unit, a refine vector may be added or subtracted from the derived motion vector. For example, the motion vector of the first prediction unit is a motion vector of the first prediction unit. It is derived by adding or subtracting the first refine vector to the first motion vector derived based on the first merge candidate, and the motion vector of the second prediction unit is the second motion vector derived based on the second merge candidate. It can be derived by adding or subtracting the refine vector. Information for determining at least one of the first refine vector or the second refine vector may be signaled through the bitstream. The information may include at least one of information for determining the size of the refine vector or information for determining the sign of the refine vector.

[406] The second refine vector may be a symmetric motion vector of the first refine vector. In this case, information for determining the only refine vector may be signaled for either of the first refine vector and the second refine vector. , When the first refine vector is determined to be (MVDx, MVDy) by the information signaled from the bitstream, the symmetric motion vector of the first refine vector (-MVDx, MVDy), (MVDx, -MVDy) or (-MVDx , -MVDy) can be set as the second refine vector. The prediction units set the symmetrical motion vector of the scaled motion vector obtained by scaling the first refine vector as the second refine vector according to the output order of each reference picture. May be

[407] As another example, one of the first prediction unit and the second prediction unit is derived based on the merge candidate, and the other motion information is signaled through the bitstream. 2020/175915 1»（：1^1{2020/002754 Can be determined based on information. For example, to signal the merge index for the first prediction unit and determine the motion vector for the second prediction unit. At least one of the information and information for determining the reference picture can be signaled. The motion information of the first prediction unit can be set equal to the motion information of the merge candidate specified by the merge index. The motion of the second prediction unit. The information may be specified by at least one of the information for determining the motion vector signaled through the bitstream and the information for determining the reference picture.

[408] Based on the motion information of the first prediction unit and the motion information of the second prediction unit, the motion prediction compensation prediction can be performed for each coding block. At this time, the image quality at the boundary of the first and second prediction units Deterioration may occur. For example, the continuity of the image quality may be lost around the edge existing at the boundary of the first prediction unit and the second prediction unit. To reduce the image quality degradation at the boundary, smoothing Predictive samples can be derived through filters or weighted predictions.

[409] The prediction sample in the coding block to which the diagonal partitioning is applied is the weighted sum operation of the first prediction sample acquired based on the motion information of the first prediction unit and the second prediction sample acquired based on the motion information of the second prediction unit. Alternatively, the prediction sample of the first prediction unit can be derived from the first prediction block determined based on the motion information of the first prediction unit, and the prediction sample of the second prediction unit is determined based on the motion information of the second prediction unit. 2 The prediction sample of the second prediction unit is derived from the prediction block, but the prediction samples located in the boundary regions of the first prediction unit and the second prediction unit are included in the first prediction sample and the second prediction block. The weighted summation operation of the included second prediction sample can be derived on the basis of the calculation. For example, Equation 3 below shows an example of inducing the prediction samples of the first prediction unit and the second prediction unit.

[410] [Equation 3]

[411] In Equation 3 above, is the first prediction sample, and P2 is the second prediction sample.

Show. wl represents the weight applied to the first predicted sample, (1-wl) represents the weight applied to the second predicted sample. As in the example shown in Equation 3, the second example, that is, the weight applied to the sample, is a constant value. In the first example, it can be derived by differentiating the weight applied to the gland.

[412] When the left triangular partition type is applied to the coding block, the boundary area may contain prediction samples having the same X-axis coordinate and y-axis coordinate. On the other hand, when the right triangular partition type is applied to the coding block, the boundary area is It is possible to include predicted samples in which the sum of the X-axis and y-axis coordinates is greater than or equal to the first threshold and less than or equal to the second threshold.

[413] The size of the boundary region is the size of the coding block, the shape of the coding block, the motion information of the prediction units, the motion vector difference value of the prediction units, the output order of the reference picture, or the first prediction sample and the second prediction sample at the diagonal border. At least one of the difference values can be determined on the basis. 2020/175915 1»（：1^1{2020/002754

32 and 33 are diagrams showing an example in which a predicted sample is derived based on a weighted sum operation of the first predicted sample and the second predicted sample. FIG. 32 is an example of a case in which the left triangular partition type is applied to the coding block. 33 illustrates a case in which the right triangular partition type is applied to the coding block. In addition, the knowledge of FIG. 32 and the 如 of FIG. 33 are diagrams showing the predicted aspects of the luma component, and FIG. 32) and FIG. 33 The (ratio of) is a drawing showing the predicted aspects of the chroma component.

[415] In the drawing, a number written in a prediction sample located near the boundary of the first prediction unit and the second prediction unit indicates a weight applied to the first prediction sample. For example, a number written in the prediction sample is In the case of, the predicted sample can be derived by applying a weight of N/8 to the first predicted sample and applying a weight of (1-uh/8)) to the second predicted sample.

[416] In the non-boundary area, the first prediction sample or the second prediction sample may be determined as a prediction sample. Referring to the example of FIG. 32, in the area belonging to the first prediction unit, based on the motion information of the first prediction unit. The derived first prediction sample may be determined as the prediction sample. On the other hand, in the region belonging to the second prediction unit, the derived second prediction sample based on the motion information of the second prediction unit may be determined as the prediction sample.

[417] Looking at the example of FIG. 33, in a region in which the sum of the X-axis coordinate and the axis coordinate is smaller than the first threshold value, the first prediction sample derived based on the motion information of the first prediction unit can be determined as the prediction sample. On the other hand, in a region in which the sum of the X-axis coordinates and the X-axis coordinates is greater than the second threshold value, a second prediction sample derived based on the motion information of the second prediction unit can be determined as the prediction sample.

[418] The threshold value for determining the non-boundary region can be determined based on at least one of the size of the coding block, the shape of the coding block, or the color component. For example, when the threshold value for the luma component is set to N, chroma The threshold for the component can be set to N/2.

[419] The predicted samples included in the boundary region can be derived based on a weighted sum operation of the first and second predicted samples. In this case, the weight applied to the first and second predicted samples is the predicted sample. It can be determined based on at least one of the location of, the size of the coding block, the shape of the coding block, or the color component.

As in the example shown in [42 as an example, in Fig. 32), the predicted samples at the same position in the X-axis coordinate and the axis coordinate can be derived by applying the same weight to the first and second prediction samples. Predicted samples with the absolute value of the difference between the X-axis and X-axis coordinates of 1 can be derived by setting the weight ratio applied to the first and second prediction samples to (3:1) or (1:3). In addition, predicted samples whose absolute value of the difference between the X-axis and X-axis coordinates is 2 can be derived by setting the weighting ratio applied to the first and second prediction samples to (7:1) or (1:7). have.

[421] Or, (as in the non-shown example, the X-axis coordinates and the axis coordinates are the same

The predicted samples of the position have the same weights for the first predicted sample and the second predicted sample. 2020/175915 1»（：1^1{2020/002754, derived by applying, and the absolute value of the difference between the X-axis coordinate and the X-axis coordinate is 1, the weighting ratio applied to the first prediction sample and the second prediction sample It can be derived by setting to (7:1) or (1:7).

[422] As an example, as in the example shown in Fig. 33), the predicted samples in which the sum of the X-axis coordinates and the axis coordinates is 1 less than the width or height of the coding block are added to the first predicted sample and the second predicted sample. It can be derived by applying the same weight. Prediction samples in which the sum of the X-axis coordinate and the X-axis coordinate equals the width or height of the coding block or are smaller than 2 are the weighting ratios applied to the first and second prediction samples (3:1) or (1:3). It can be derived by setting it to ). The sum of the X-axis coordinate and X-axis coordinate is less than the width or height of the coding block.

Estimated samples that are larger than 1 or smaller than 3 can be derived by setting the weight ratio applied to the first and second predicted samples to (7:1) or (1:7).

[423] Or, (as in the non-shown example, the predicted samples in which the sum of the X-axis coordinates and the axis coordinates is 1 less than the width or height of the coding block are the same as those of the first predicted sample and the second predicted sample. It can be derived by applying a weight.For the predicted samples where the sum of the X-axis coordinate and the X-axis coordinate is equal to the width or height of the coding block or less than 2, the weighting ratio applied to the first predicted sample and the second predicted sample is (7: It can be derived by setting it to 1) or (1:7).

[424] As another example, a weight value can be determined taking into account the position of the predicted sample or the shape of the coding block. Equations 4 to 6 are examples of inducing the weight when the left triangular partition type is applied to the coding block. Equation 4 shows an example of deriving a weight applied to the first predicted sample when the coding block is square.

[425] [Equation 4]

[426] In Equation 4, X and denote the position of the prediction sample. When the coding block is amorphous, the weight applied to the first prediction sample can be derived as in Equation 5 or Equation 6. 5 shows the case where the width of the coding block is greater than the height, and Equation 6 shows the case where the width of the coding block is less than the height.

[427] [Equation 5]

>1 _* 1 =( (c/ /， ¾¾" ?)- ₎ +4)/8

[428] [Equation 6]

[429] When the right triangular partition type is applied to the coding block, the weight applied to the first predicted sample can be determined as in Equation 7 to Equation 9. Equation 7 is the first predicted sample when the coding block is square. It shows an example of deriving the weights applied to it.

[430] [Equation ⁷ ]

wl ⁼ (Ja 1 -j·> )+4)/8

[431] In Equation 7, [¾ represents the width of the coding block. The coding block is non-square 2020/175915 1»（：1^1{ In the case of 2020/002754, the weight applied to the first prediction sample can be derived as in Equation 8 or Equation 9. Equation 8 shows that the width of the coding block is greater than the height. A large case is indicated, and Equation 9 indicates a case where the width of the coding block is smaller than the height.

[432] [Equation 8]

[433] [Equation 9]

In Equation 8, 幻^ represents the height of the coding block.

[435] As in the illustrated example, the ones included in the first prediction unit among the prediction samples within the boundary region are derived by giving a greater weight to the first prediction sample than the second prediction sample, and included in the second prediction unit. This can be derived by giving a greater weight to the second predicted sample than to the first predicted sample.

[436] When diagonal partitioning is applied to a coding block, a combined prediction mode in which an intra prediction mode and a merge mode are combined may be set not to be applied to the coding block.

[437] When the coding/decoding of the coding block is completed, motion information of the coding/decoding completed coding block can be stored for the coding/decoding of the next coding block.

The motion information may be stored in units of subblocks having a preset size. For example, a subblock having a preset size may have a size of 4x4. Or, depending on the size or shape of a coding block, the size or shape of the subblock It can be determined virtually.

[438] When the subblock belongs to the first prediction unit, the movement information of the first prediction unit can be stored as the movement information of the subblock. On the other hand, when the subblock belongs to the second prediction unit, Motion information can be saved as sub-block motion information.

[439] When the sub-block spans the boundary of the first and second prediction units, one of the motion information of the first prediction unit and the motion information of the second prediction unit can be set as the motion information of the sub-block. For example, the motion information of the first prediction unit may be set as the motion information of the sub-block, or the motion information of the second prediction unit may be set as the motion information of the sub-block.

움직임정보및 움직임 정보중다른 하나는제 2예측유닛으로부터유도할수있다.일 예로,제 1예측유닛의 ) 움직임 정보를서브블록의 )움직임 정보로설정하고,제 2예측유닛의 1그 움직임 정보를서브블록의

움직임 정보로설정할수있다.다만,제 1예측 유닛및제 2예측유닛이

움직임정보만을갖거나, 1그움직임정보만을갖는 경우,제 1예측유닛또는제 2예측유닛중어느하나를선택하여,서브블록의 움직임 정보를결정할수있다.또는,제 1예측유닛및제 2예측유닛의움직임 2020/175915 1»（：1^1{2020/002754 벡터 평균값을서브블록의움직임 벡터로설정할수있다. [44] As another example, when the sub-block crosses the boundary of the first prediction unit and the second prediction unit, one of the movement information and the kinematic information of the sub-block is derived from the first prediction unit, and

The other one of the motion information and the motion information can be derived from the second prediction unit. For example, the motion information of the first prediction unit is set as the motion information of the subblock, and the motion information of the second prediction unit is set. Subblock

It can be set by motion information, but the first prediction unit and the second prediction unit

In the case of having only the motion information or only the first motion information, the motion information of the sub-block can be determined by selecting either the first prediction unit or the second prediction unit. Or, the first prediction unit and the second prediction unit Unit movement 2020/175915 1»（：1^1{2020/002754 The vector average value can be set as the motion vector of the subblock.

[441] The motion information of the coding block after encoding/decoding is completed is stored in the motion information table.

At this time, the motion information of the coding block to which the prediction unit partitioning has been applied can be set not to be added to the motion information table.

[442] Alternatively, it is possible to add motion information of only one of the plurality of prediction units generated by dividing the coding block to the motion information table. As an example, the motion information of the first prediction unit is added to the motion information table. ,The motion information of the second prediction unit may not be added to the motion information table. At this time, the size of the coding block, the shape of the coding block, the size of the prediction unit, the shape of the prediction unit, or whether the two-way prediction was performed for the prediction unit. Based on at least one of whether or not, a prediction unit to be added to the motion information table can be selected.

[443] Alternatively, motion information of each of a plurality of prediction units generated by dividing a coding block may be added to the motion information table. At this time, the order of adding the motion information table may be predefined in the encoder and decoder. For example, motion information of a prediction unit including an upper left sample or a lower left corner sample may be added to the motion information table prior to the motion information of other prediction units. Or, merge index, reference picture index, or motion information of each prediction unit. The order of addition to the motion information table can be determined based on at least one of the sizes of the vector.

[444] Alternatively, motion information obtained by combining the motion information of the first prediction unit and the motion information of the second prediction unit can be added to the motion information table. Either of the motion information and the motion information of the combined motion information It is derived from the first prediction unit, and 0) the other of the motion information and the knee motion information may be derived from the second prediction unit.

[445] Or, whether the reference pictures of the first prediction unit and the second prediction unit are the same

Based on the motion information to be added to the motion information table, motion information to be added to the motion information table can be determined. For example, when the reference pictures of the first and second prediction units are different, the motion information of any one of the first and second prediction units Alternatively, motion information obtained by combining the first prediction unit and the second prediction unit can be added to the motion information table. On the other hand, when the reference pictures of the first prediction unit and the second prediction unit are the same, the motion vector and the second prediction unit 2 The average of the motion vectors of the prediction unit can be added to the motion information table.

[446] Or, according to the size of the coding block, the shape of the coding block, or the division type of the coding block.

Based on the motion vector, the motion vector to be added to the motion information table can be determined. For example, when right triangular partitioning is applied to the coding block, the motion information of the first prediction unit can be added to the motion information table. On the other hand, in the coding block. When the left triangular partitioning is applied, the motion information of the second prediction unit can be added to the motion information table, or motion information obtained by combining the motion information of the first prediction unit and the motion information of the second prediction unit can be added to the motion information table. . 2020/175915 1»（：1^1{2020/002754

[447] A motion information table for storing motion information of a coding block to which prediction unit partitioning is applied may be separately defined. For example, motion information of a coding block to which prediction unit partitioning is applied may be stored in a partition mode motion information table. The split mode motion information table can also be referred to as a triangular motion information table, i.e., the motion information of the coding block to which the prediction unit partitioning is not applied is stored in the general motion information table, and the motion information of the coding block to which the prediction unit partitioning is applied is Examples of adding the motion information of a coding block to which the above-described prediction unit partitioning is applied to the motion information table can be applied to updating the split mode motion information table. For example, the split mode motion information table can be stored in the split mode motion information table. In the motion information table, the motion information of the first prediction unit, the motion information of the second prediction unit, the motion information combining the motion information of the first prediction unit and the motion information of the second prediction unit, and the motion vector task of the first prediction unit. 2 Motion information obtained by averaging the motion vector of the prediction unit can be added.

[448] When prediction mode partitioning is not applied to a coding block, a merge candidate can be derived using a general motion information table. On the other hand, when prediction mode partitioning is applied to a coding block, merge using the prediction mode motion information table Can induce candidates.

[449]

[45 When the merge candidate of the current block is selected, the motion vector of the selected merge candidate is set as the initial motion vector, and the motion vector derived by adding or subtracting the offset vector to the initial motion vector is used to move to the current block. Compensation prediction can be performed. A merge offset vector encoding method can be defined to induce a new motion vector by adding or subtracting the offset vector to the motion vector of the merge candidate.

[451] Information indicating whether to use the merge offset encoding method

It can be signaled through a bitstream. The information is a 1-bit flag.

블록에 머지오프셋벡터 부호화방법이 적용되지 않음을나타낸다.머지오프셋부호화방법이 적용되지 않는경우,머지후보의움직임 벡터가현재블록의움직임 벡터로설정될수 있다. 111 can be 6_（none 861;_¥6。1;0]'_£1&silver. For example, 1116 6_(none 861;_¥6。1;0]'_£1& silver is 1). Indicates that the merge offset vector encoding method is applied to the current block. When the merge offset vector encoding method is applied to the current block, the motion vector of the current block can be derived by adding or subtracting the offset vector to the motion vector of the merge candidate. .

This indicates that the merge offset vector encoding method is not applied to the block. When the merge offset encoding method is not applied, the motion vector of the merge candidate may be set as the motion vector of the current block.

블록에머지모드가적용됨을 2020/175915 1»（：1^1{2020/002754 나타내는 1116 6_：（¾은의 값이 1인경우, 111 은6_0 861；_¥6아；0]'_£^'^가부호화되어 시그날링될수있다. [452] The flag can be signaled only when the value of the skip flag indicating whether the skip mode is applied is true or the value of the merge flag indicating whether the merge mode is applied is true. For example, the current block Indicating that the skip mode is applied

Block Merge Mode is applied. 2020/175915 1»（：1^1{2020/002754 indicates 1116 6_：（when the value of ¾ is 1, 111 is 6_0 861;_¥6ah;0]'_£ ^' ^ can be encoded and signaled have.

[453] When it is determined that the merge offset encoding method is applied to the current block, at least one of information specifying one of the merge candidates included in the merge candidate list, information indicating the size of the offset vector, or information indicating the direction of the offset vector One can be signaled additionally.

[454] Information for determining the maximum number of merge candidates that the merge candidate list can contain can be signaled through the bitstream. For example, the maximum number of merge candidates that the merge candidate list can contain is set to a natural number of 6 or less. Can be

[455] When it is determined that the merge offset encoding method is applied to the current block, only the preset maximum number of merge candidates can be set as the initial motion vector of the current block. That is, depending on whether the merge offset encoding method is applied or not, The number of merge candidates for which the current block is available can be adaptively determined.

If the value of is set to 0, the maximum number of merge candidates that the current block can use is set to ], while 111 6_0 861;_¥ 1;0]'_1¾ is set to 1, the current block can use. The maximum number of merge candidates can be set to N, where M represents the maximum number of merge candidates that the merge candidate list can contain, and N represents a natural number equal to or less than M.

[456] For example, if M is 6 and N is 2, the two merge candidates with the smallest index among the merge candidates included in the merge candidate list may be set as available for the current block. Accordingly, the index The motion vector of the merge candidate with a value of 0 or the motion vector of the merge candidate with the index value of 1 can be set as the initial motion vector of the current block. If M and N are the same (e.g., if M and N are 2) ）, All merge candidates included in the merge candidate list can be set as available for the current block.

[457] Or, whether a neighboring block is available as a merge candidate is merged into the current block.

It can be determined based on whether or not the offset vector encoding method is applied. For example, if the value of 111 6_0^61;_¥ 1;0]'_£ ^' ^ is 1, it is adjacent to the right, that is, the top corner of the current block. At least one of a neighboring block, a neighboring block adjacent to the lower left corner, or a neighboring block adjacent to the lower left corner may be set to be unavailable as a merge candidate. Accordingly, if the merge offset vector coding method is applied to the current block, the current block At least one of the neighboring blocks adjacent to the right or upper corner of the block, the neighboring block adjacent to the left or lower corner, or the neighboring block adjacent to the lower left corner cannot be set as the initial motion vector, or 1116 _0^61; If the value of _¥6;0]'_£ ^' ^ is 1, the temporal neighboring block of the current block may be set as unavailable as a merge candidate.

경우,머지후보리스트에포함된 머지후보들의 개수가최대개수보다작은경우라하더라도,페어와이즈머지 후보또는제로머지후보중적어도하나는머지후보리스트에추가되지 않을 수있다. [458] When the merge offset vector encoding method is applied to the current block, pairwise merge 2020/175915 1»（：1^1{2020/002754 Candidate or zero merge candidate can be set not to use at least one. Accordingly,

In this case, even if the number of merge candidates included in the merge candidate list is less than the maximum number, at least one of the pairwise merge candidates or zero merge candidates may not be added to the merge candidate list.

[459] The motion vector of the merge candidate can be set as the initial motion vector of the current block. At this time, when the number of merge candidates that the current block can use is plural, information specifying any one of the plural merge candidates may be signaled through the bitstream. For example, the maximum number of merge candidates that the merge candidate list can include may be signaled. When the number is greater than 1, information indicating any one of the plurality of merge candidates can be signaled through the bitstream, i.e., under the merge offset encoding method, any one of the plurality of merge candidates can be signaled through the bitstream.

Candidate can be specified in the current block

It can be set as the motion vector of the merge candidate.

있다.즉,머지오프셋 부호화방법하에서,머지후보리스트에 1개의머지후보가포함되어 있을경우, 머지후보를특정하기 위한정보 의부호화를생략하고,머지후보 리스트에포함된머지후보를기초로,초기움직임 벡터를결정할수있다.상기 머지후보의움직인벡터가현재블록의초기움직임 벡터로설정될수있다. [On the other hand, if the number of merge candidates that the current block can use is 1, the signaling of information to specify the merge candidate can be omitted. For example, the maximum number of merge candidates that the merge candidate list can include is 1 If not greater than,

In other words, under the merge offset coding method, if one merge candidate is included in the merge candidate list, the encoding of information for specifying the merge candidate is omitted, and based on the merge candidate included in the merge candidate list, the initial The motion vector can be determined. The motion vector of the merge candidate can be set as the initial motion vector of the current block.

있다.표[461] As another example, after determining the merge candidate for the current block, it is also possible to decide whether to apply the merge offset vector encoding method to the current block. For example, the maximum number of merge candidates that a merge candidate can contain is less than 1. In large cases, information for specifying one of the merge candidates 111 6_:1(1 can be signaled. After selecting the merge candidate based on 1116 _:1(1), the merge offset vector encoding method is applied to the current block. Indicating whether or not

There is.

3 is a diagram showing a syntax table according to the above-described embodiment.

[462] [S.5]

2020/175915 1»(：1^1{2020/002754

보다큰경우, [463] As another example, after determining the merge candidate of the current block, the merge offset vector encoding method in the current block only when the index of the determined merge candidate is less than the maximum number of merge candidates that can be used when the merge offset vector encoding method is applied. It is also possible to decide whether to apply the merge offset vector encoding method to the current block, for example, only when the value of the index information is smaller.

Greater than,

의값이 N보다 2020/175915 1»（：1^1{2020/002754 작고,상기 인덱스정보에 의해선택된머지후보가양방향움직임 정보를갖는 경우에 한하여,현재블록에 머지오프셋벡터부호화방법을적용할것인지 여부를나타내는 1116 6_(났861；_¥6。1；0]'_£1&은를부호화하여시그날링할수있다. 또는,인덱스정보 산 의 값이 보다작고,상기 인덱스정보에의해 선택된머지후보가단방향움직임 정보를갖는경우에 한하여,현재블록에 머지오프셋벡터부호화방법을적용할것인지 여부를나타내는 111 Silver 6_（Look 861；_¥6。1；0]'_£1& silver encoding may be omitted. If the encoding of 111 is 6_0 861;_¥6。1;0]'_1¾ is omitted, it may be determined that the merge offset vector encoding method is not applied to the current block. Or, after determining the merge candidate of the current block, Considering whether the determined merge candidate has bidirectional motion information or unidirectional motion information, it is possible to decide whether to apply the merge offset vector encoding method to the current block. For example, index information

Is less than N 2020/175915 1»（：1^1{2020/002754 1116 indicating whether or not to apply the merge offset vector encoding method to the current block only when the merge candidate selected by the index information is small and has bidirectional motion information. 6_(Now 861;_¥6。1;0]'_£1&) can be signaled by encoding or, the value of the index information mountain is smaller, and the remaining candidate selected by the index information has one-way motion information. In this case, it indicates whether to apply the merge offset vector encoding method to the current block.

111 can be signaled by encoding 6_(none861;_¥6。1;0]'_典&silver.

[464] Or, the size, shape, or current block of the current block is at the boundary of the coding tree unit.

Based on at least one of whether or not it is in contact, it is possible to decide whether to apply the merge offset vector encoding method. At least one of the size and shape of the current block or whether the current block is in contact with the boundary of the coding tree unit is a preset condition. If is not satisfied, the encoding of, indicating whether to apply the merge offset vector encoding method to the current block, can be omitted.

[465] When a merge candidate is selected, the motion vector of the merge candidate can be set as the initial motion vector of the current block. Then, information indicating the size of the offset vector and information indicating the direction of the offset vector are decoded to obtain the offset vector. I can decide. The offset vector can have a horizontal direction component or a vertical direction component.

있다.표[466] The information indicating the size of the offset vector may be index information indicating any one of the vector size candidates. For example, indicating any one of the vector size candidates.

There is.

4 represents the value of the variable DistFromMergeMV for determining the size of the offset vector according to the index information (1 &11.6_:1(Urine binarization and (1 &1 6_:1(1).

[467] [Table 6]

[468] The size of the offset vector can be derived by dividing the variable DistFromMergeMV by a preset value. The equation shows an example of determining the size of the offset vector. 2020/175915 1»（：1/10公020/002754

[469] [Equation 1

[If 47 follows Equation 10 above, the variable DistFromMegeMV is divided by 4 or the variable

You can set the value of DistFromMergeMV shifted by 2 bits to the left as the size of the offset vector.

[471] It is also possible to use a larger number of vector size candidates or a smaller number of vector size candidates than the example shown in Table 6, or to set the range of motion vector offset size candidates different from those shown in Table 6. For example, It can be set so that the size of the horizontal or vertical component of the offset vector is no larger than 2 sample distances. Table 7 shows the index information (1 &11.6_:1 (binaryization of urine).

Represents the value of the variable DistFromMergeMV to determine the size of the vector.

[472] [Table 7]

[473] Or, based on the motion vector precision, the motion vector offset size candidates

The range can be set differently. For example, if the motion vector precision for the current block is fractional-pel, the values of the variable DistFromMergeMV corresponding to the values of the index information distance_idx are 1, 2, 4, 8, 16. It can be set to etc. Here, the minority pel includes at least one of 1/16 pel, Octo-pel, Quarter-pel, or Half-pel, while movement relative to the current block. When the vector precision is an integer-pel, the values of the variable DistFromMergeMV corresponding to the values of the index information distance_idx can be set to 4, 8, 16, 32, 64, etc. That is, according to the motion vector precision for the current block. ,variable

In order to determine DistFromMergeMV, more than the referenced table can be set. For example, if the motion vector precision of the current block or merge candidate is a quarterfel, then Table 6 can be used to derive the variable DistFromMergeMV pointed to by distance_idx. If the motion vector precision of the block or merge candidate is an integer pel, the value obtained by taking N times (e.g., 4 times) to the value of the variable DistFromMergeMV indicated by distance_idx in Table 6 can be derived as the value of the variable DistFromMergeMV.

[474] Information for determining the motion vector precision may be signaled through a bitstream. For example, the information may be signaled at the sequence, picture, slice or block level. Accordingly, the range of vector size candidates is , Via bitstream 2020/175915 1»（：1^1{2020/002754 It can be set differently by the information related to the motion vector precision signaled. Alternatively, the motion vector precision can be determined based on the merge candidate of the current block. For example, the motion vector precision of the current block may be set equal to the motion vector precision of the merge candidate.

[475] Alternatively, information for determining the search range of the offset vector can be signaled through a bitstream. Based on the search range, at least one of the number of vector size candidates and the minimum or maximum of the vector size candidates can be determined. For example, to determine the search range of the offset vector, there is a neat 1116 6_(^861；_¥ 1；0]'_£ ^' ^

It can be signaled through a bitstream. The information can be signaled through a sequence header, a picture header, or a slice header.

벡터의크기가[476] For example,

The size of the vector

대값은 128로설정될수있다. It can be set not to exceed 2. Accordingly, the maximum value of DistFromMergeMV can be set to 8. On the other hand, 111 is 6_0 861;_6 Yes 611 (1_ 1¾6_: (If the value is 1, the offset vector size is 32 samples) It can be set so as not to exceed the distance.

The large value can be set to 128.

[477] The size of the offset vector can be determined by using a flag indicating whether the size of the offset vector is greater than the threshold value. For example, the size of the offset vector is

Flag indicating whether it is greater than the threshold

The threshold can be 1, 2, 4, 8 or 16. For example,

A value of 1 indicates that the size of the offset vector is greater than 4, while a value of 0 indicates that the size of the offset vector is less than 4.

의크기가문턱값보다

[478]

Size is less than the threshold

:1(뇨의 부호화양상을나타낸신택스테이블이다. Using, can derive the difference between the size of the offset vector and the threshold value. Or, if the size of the offset vector is less than or equal to the threshold, the size of the offset vector can be determined using Table 8

:1 (This is a syntax table showing the coding pattern of urine.

[479] [S.8]

2020/175915 1»(：1/10公020/002754

[480] Equation 11 shows an example of inducing the variable DistFromMergeMV for determining the size of the offset vector using distance_flag and distance_idx.

[481] [Equation 11]

DistFromMergeMV = N ^ distance flag + (1 «distance _idx)

In Equation 11, the value of distance_flag may be set to 1 or 0.

The value of distance_idx can be set to 1, 2, 4, 8, 16, 32, 64, 128, etc. N represents the coefficient determined by the threshold value. For example, if the threshold value is 4, N can be set to 16.

[483] The information indicating the direction of the offset vector may be index information indicating any one of the vector direction candidates. For example, any one of the vector direction candidates may be used. 2020/175915 1»（：1^1{2020/002754 Indicative index information scene 0!1_：1（Urine can be signaled through the bitstream. Table 9 shows index information（1 urine 6（No.011_：1（ Binarization of urine and (1 urine 6 (number 011_: 1) indicates the direction of the offset vector according to urine.

[484] [Table 9]

[485] In Table 9, [This indicates the horizontal direction, and 四⑴ indicates the vertical direction. +1 indicates that the value of the X component or X component of the offset vector is +, and -1 indicates that the value of the X component or component of the offset vector is -. Give an example of determining

[486] [Equation 12]

_< ¾分 1넜7 <ίJ^）*s^gn [0]

（I bought 7: ^/ ） * _5/玄₇₁ [1]

[이는오프셋벡터의수직방향성분을나타내고, offsetMV[l]은오프셋벡터의수평 방향성분을나타낸다. [487] Equation

[This represents the vertical direction component of the offset vector, and offsetMV[l] represents the horizontal direction component of the offset vector.

방향을 나타내는（1뇨6 011_：1（1 의 값에따른오프셋벡터를나타낸도면이다. [488] Figure 34 shows the size of the offset vector

This is a diagram showing the offset vector according to the value of (1 urine 6 011_: 1 (1) indicating the direction.

[489] As in the example shown in FIG. 34, the size and direction of the offset vector is

Sin^ 011_:1 (Can be determined according to the value of urine. The maximum size of the offset vector can be set so as not to exceed the threshold value. Here, the threshold value can have a predefined value in the encoder and the decoder. Thus, the threshold value may be 32 sample distances. Alternatively, depending on the size of the initial motion vector, the threshold value may be determined. For example, the threshold value for the horizontal direction is based on the size of the horizontal component of the initial motion vector. Is set, and the threshold value for the vertical direction may be set based on the size of the vertical direction component of the initial motion vector.

움직임 벡터를 현재블록의 ）초기움직임 벡터로설정하고,머지후보의니움직임 벡터를 현재블록의 1초기움직임 벡터로설정할수있다.이때,머지후보의 1_^0참조 픽처와현재픽처와의출력순서차분값（이하, 0차분값이라함）과머지후보의 참조픽처와현재픽처와의출력순서차분값（이하, 1 차분값이라함）을 고려하여, 오프셋벡터 및니오프셋벡터를결정할수있다. [49 If the merge candidate has bidirectional movement information, the merge candidate's

The motion vector can be set as the initial motion vector of the current block, and the motion vector of the merge candidate can be set as the initial motion vector of the current block. At this time, the output sequence difference between the 1 _^ 0 reference picture of the merge candidate and the current picture The offset vector and the knee offset vector can be determined by taking into account the value (hereinafter referred to as the zero difference value) and the output order difference value (hereinafter referred to as the first difference value) between the reference picture and the current picture of the remaining candidate.

부호화동일한경우,

오프셋벡터와

[491]

If the encoding is the same,

Offset vector and

The offset vector can be set the same, whereas the sign of the first difference value and the tooth difference value is 的

2020/175915 1»（：1^1{2020/002754 When different: The knee offset vector can be set in the opposite direction to the inner offset vector.

[492] The size of the inner offset vector and the size of the offset vector can be set equally. Or, it is possible to determine the size of the knee offset vector by scaling the offset vector based on the difference value and the tooth difference value.

[493] As an example, Equation 13 shows: ）when the sign of the difference value and the tooth difference value are the same, ）the offset vector and 1 _^ 1 offset vector are represented.

[494] [Equation 13]

오프셋 벡터의수직방향성분을나타낸다. [495] In Equation 13, offsetMVL() [This represents the horizontal direction component of the inner offset vector, and offsetMVL0[l] represents the vertical direction component of the 1-刀 offset vector. offsetMVLl [This represents the horizontal direction component of the offset vector, and offsetMVLl[l] is

Represents the vertical component of the offset vector.

[496] Equation 14 shows the inner offset vector and the knee offset vector when the signs of the inner difference value and the difference value are different.

[497] [Equation 14]

[498] It is also possible to define more than four vector-direction candidates. Tables and Tables 11 show examples in which eight vector-direction candidates are defined.

[499] [Table 1]

2020/175915 1»（：1^1{2020/002754

2020/175915 1»（：1^1{2020/002754

[500] [Table 11]

[501] Silver in Table and Table 11!! [This and the absolute value of 11[1] greater than 0 indicates that the offset vector is in a diagonal direction. When Table 9 is used, the size of the X-axis and axis component of the diagonal offset vector is set to abs (offsetMV). On the other hand, when Table 10 is used, the sizes of the X-axis and X-axis components of the diagonal offset vector can be set to abs (offsetMV/2).

Fig. 35 is a diagram showing the offset vector according to the value of (1 urine 6 011_: 1 (1) indicating the size of the offset vector and the direction of the offset vector.

[503] In FIG. 35 is an example of a case where Table 9 is applied, and in FIG.

This is an example of the case applied.

트스트림을통해시그날링될 수있다.상기플래그는시퀀스,픽처또는슬라이스레벨에서시그날링될수 있다.일 예로,상기플래그의 값이 0인경우,표 9에 예시된 4개의 벡터방향 후보들을이용할수있다.반면,상기플래그의 값이 1인경우,표 또는표 11에 예시된 8개의 벡터 방향후보들을이용할수있다. [504] Information for determining at least one of the number or size of vector direction candidates may be signaled through the bitstream. For example, determining vector direction candidates

The flag may be signaled through a stream. The flag may be signaled at a sequence, picture, or slice level. For example, when the value of the flag is 0, the four vector direction candidates illustrated in Table 9 can be used. On the other hand, if the value of the flag is 1, the 8 vector direction candidates illustrated in Table or Table 11 can be used.

[505] Alternatively, based on the size of the offset vector, at least one of the number or size of vector direction candidates can be determined. For example, the value of the variable DistFromMergeMV for determining the size of the offset vector is equal to or equal to the threshold. If the value is smaller than the value, the eight vector direction candidates illustrated in Table or Table 11 can be used, whereas if the value of the variable DistFromMergeMV is greater than the threshold value, the four vector direction candidates illustrated in Table 9 can be used.

[506] Alternatively, based on the value of the X component MVx and the value of the X component MVy of the initial motion vector, at least one of the number or size of vector direction candidates can be determined. For example, MVx and MVy difference or absolute difference of the difference If the value is below the threshold, Table 10 or Table

The eight vector direction candidates illustrated in 11 are available, whereas the MVx and MVy differences 2020/175915 1»（：1^1{2020/002754 or if the absolute value of the difference is greater than the threshold, the four vector direction candidates shown in Table 9 can be used.

[507] The motion vector of the current block can be derived by adding the offset vector to the initial motion vector. Equation 15 shows an example of determining the motion vector of the current block.

[508] [Equation 15]

[509] In Equation 15, 11110 denotes the inner motion vector of the current block, and 11111 denotes the kinematic vector of the current block. mergeMVL0 represents the initial motion vector of the current block (that is, the motion vector of the merge candidate), and mergeMVLl represents the initial motion vector of the current block. [This represents the horizontal direction component of the motion vector, and [1] represents the vertical direction component of the motion vector.

[510]

[511] The affine seed vector derived based on the affine merge mode or the afine motion vector prediction mode, or the motion vector of a subblock (subblock motion vector or afine subblock vector) can be updated based on the offset vector. Specifically, an offset is added to the affineseed vector or the motion vector of the subblock, or

By subtracting, the updated affineseed vector or the updated sub-block motion vector can be derived Under the affine motion model, refining the affineseed vector or the sub-block motion vector can be referred to as the afine merge offset encoding method. have.

[512] If the afine motion model is applied to the coding block and the value of the lag, indicating whether the merge offset coding method is used, is 1, the afine merge offset coding method can be applied to the coding block.

[513] If it is determined that the affine merge offset coding method is applied,

복수의크기후보들중어느하나를가리키고, 방향을결정하기위한인덱스정보（1뇨6 011_：1（1 는복수의방향후보들중어느 하나를가리킨다.

방향을 결정하기 위한인덱스정보（1 011_:1（1 에기초하여,오프셋벡터（（났요 쇼 노아이, 0 8 쇼&1아1]）가결정될수있다. The merge index for determining the initial motion vector of the block 知16¾6_：1 (urea), the index information for determining the size of the offset vector, and the index information for determining the direction of the offset vector. Can be determined size

Points to any one of the plurality of size candidates, and index information for determining the direction (1 urine 6 011_: 1 (1 indicates any one of the plurality of direction candidates).

The index information for determining the direction (1 011_:1 (based on 1, offset vector ((None Sho noai, 0 8 Sho & 1A1))) can be determined.

참조픽처의시간적방향이동일한것으로결정될수있다. [514] The updated affine seed vector can be derived by adding or subtracting the offset vector to the affine seed vector. At this time, the sign of the offset vector applied to each affine seed vector can be determined according to the direction of the reference picture. For example, in the coding block 2020/175915 1» (：1^1{2020/002754 bidirectional prediction is applied,) When the temporal direction of the reference picture and the temporal direction of the reference picture are the same, as shown in Equation 16 below, each word fineseed vector An offset vector can be added to the offset vector. Here, the temporal direction can be determined based on the output order of the current picture and the reference picture 00. For example, the output order difference between the current picture and the reference picture, and the current picture and reference. If the output order difference of pictures is all negative, or if the output order difference between the current picture and my reference picture and the output order difference between the current picture and this picture are both positive, the temporal direction of the reference picture and

The temporal direction of the reference picture can be determined to be the same.

[515] [Equation 16]

[516] On the other hand, when the temporal direction of 1,0 reference pictures and knee reference pictures is different, the following

As shown in Equation 17, an offset vector is added to each word fineseed vector or

차분은음수이나현재픽처와니참조픽처의 출력순서차분은양수인경우,또는현재픽처와 0참조픽처의출력순서 차분은양수이나현재픽처와 참조픽처의출력순서차분은음수인경우, ) 참조픽처의시간적방향과 1_^1참조픽처의시간적방향이상이한것으로결정될 수있다. By subtracting, an updated affineseed vector can be derived. For example, the current

If the difference is negative, but the difference in the output order of the current picture and the reference picture is positive, or the difference in the output order of the current picture and the reference picture 0 is positive, but the difference in the output order of the current picture and the reference picture is negative, the temporal difference of the reference picture It can be determined that the direction and the temporal direction of the 1 _^ 1 reference picture are the same.

[517] [Equation 17]

[518] In Equation 16 and Equation 17, the same offset for all affineseed vectors

Although the vector is illustrated as being applied, the present invention is not limited to this. It is also possible to determine the offset vector of each affine vector individually.

[519] Alternatively, an offset vector can be set for each subblock. The motion vector of a subblock can be updated using the offset vector of the subblock.

[52 depends on the motion vector precision of the current block or the neighboring block, the offset vector size

The range of candidates can be determined differently, i.e., depending on the motion vector precision of the current block or neighboring block, the number of candidates for the offset vector size, minimum or maximum value. 2020/175915 1»（：1^1{2020/002754 At least one may be different. For example, if the afine motion model is applied to the current block, and the motion vector precision of the current block is 1/4 pel, the variable

의해 1, 2, 4, 8, 16, 32, 64및 128중어느 하나로결정될수있다.반면,현재블록의움직임 벡터 정밀도가정수펠인경우, 변수 DistFromMergeMV의

의해 4, 8, 16, 32, 64, 128, 256및DistFromMergeMV

It can be determined to be one of 1, 2, 4, 8, 16, 32, 64, and 128, whereas if the current block's motion vector precision is an integer pel, the variable DistFromMergeMV

By 4, 8, 16, 32, 64, 128, 256 and

It can be determined as any one of 512.

[521] As another example, information for specifying any one of a plurality of offset vector size candidate sets may be signaled through the bitstream. The number of offset vector size candidates each of the offset vector size candidate sets contains, or At least one of the types may be different. For example, when the first offset vector size candidate set is selected, the variable DistFromMergeMV may be determined as one of {1, 2, 4, 8, 16, 32, 64, 128 }, and 2 When the offset vector size candidate set is selected, the variable DistFromMergeMV can be determined as one of {4, 8, 16, 32, 64, 128, 256, 512}.

[522] Index information DistMV_idx specifying any one of a plurality of offset vector size candidate sets may be signaled through a bitstream. For example, DistMV_idx of 0 indicates that the first offset vector size candidate set is selected,

If DistMV_idx is 1, it indicates that the second offset vector size candidate set is selected.

[523]

[524] It is also possible to set the offset vector for each subblock or sample. That is, subblocks

Or offset vector (or difference vector) or offset vector for samples

An array (or difference vector array) can also be defined as offset data.

[525] For example, if a subblock motion vector is derived based on affineseed vectors, motion compensation for the subblock can be performed using the induced subblock motion vector. At this time, the subblock or the subblock or the subblock when performing the motion compensation is performed. Additional sample-specific offset vectors are available.

[526] The offset vector for the subblock can be derived using the offset vector candidate. Based on the offset vector, it is possible to update the sub-block motion vector, and perform motion compensation for the sub-block based on the updated sub-block motion vector.

[527] An offset vector for each prediction sample within a subblock may be derived. Specifically, based on the position of each prediction sample within the subblock, an offset vector for each prediction sample can be derived. Here, the location of the prediction sample is It can be determined based on the upper left sample of the subblock.

[528] The X component of the offset vector for the predicted sample is the difference between the X component of the second affineseed vector and the X component of the first affineseed vector, multiplied by the X-axis coordinate of the predicted sample, and the second affineseed. It can be derived based on the difference between the vector X component and the first affineseed vector component multiplied by the X-axis coordinate of the predicted sample. The offset vector component for the predicted sample is the third affineseed vector X component. And the first affineseed vector X component 2020/175915 1»（：1^1{2020/002754 The difference value multiplied by the X-axis coordinate of the prediction sample, and the difference value between the third affineseed vector component and the second affineseed vector X component, and the predicted sample It can be derived based on the product of the axial coordinate of

[529] When a 4-parameter motion model is applied to the current block, the component of the offset vector is the X-axis coordinate of the predicted sample to the difference between the first affineseed vector X component and the second affineseed vector X component. A value obtained by multiplying the product by multiplying the axial coordinates of the predicted sample by the difference between the second affine seed vector component and the first affine seed vector X component can be derived.

As [53] described above, the offset vectors of the prediction samples in the sub-block may have different values. However, the offset vector array for the prediction samples can be commonly applied to all sub-blocks. That is, the first sub-block. The offset vector array applied to the block and the offset vector array applied to the second subblock may be the same.

[531] Alternatively, an offset vector array for each sample can be derived by taking the position of the subblock into consideration. In this case, an offset vector array different between the subblocks can be applied.

[532] After performing motion compensation for the subblock based on the subblock motion vector, each prediction sample can be updated based on the offset vector. The update of the prediction sample is performed on the offset vector of the prediction sample and the prediction sample. Can be performed on the basis of gradients.

[533] The gradient for the predicted sample can be derived based on the difference value of the predicted samples. The gradient for the first predicted sample is the difference value between the first predicted sample and the predicted samples belonging to the same line, or the first predicted sample. It can be derived based on the difference value between the prediction samples belonging to the line adjacent to.

[534] As an example, the gradient for the first predicted sample may be derived as a difference value between the first predicted sample and another predicted sample belonging to the same line as the first predicted sample. Specifically, the horizontal direction gradient of the first predicted sample Is derived as the difference value between the first predicted sample and the second predicted sample belonging to the same row as the first predicted sample, and the vertical gradient of the first predicted sample belongs to the same column as the first predicted sample and the first predicted sample. 3 can be derived as a difference value from the predicted sample. Here, the second predicted sample and the third predicted sample may be adjacent to the first predicted sample. For example, the second predicted sample is to the left or right of the first predicted sample. And the third predicted sample may be positioned above or below the first predicted sample; or, the second predicted sample and the third predicted sample are spaced a predetermined distance from the first predicted sample in the X-axis and/or X-axis direction. Here, the predetermined distance may be a natural number such as 1, 2 or 3.

[535] Alternatively, the difference value of the predicted samples belonging to the line adjacent to the first predicted sample may be set as the gradient with respect to the first predicted sample. For example, the horizontal direction gradient with respect to the first predicted sample is the first predicted sample. It can be derived as a difference value of the predicted samples belonging to the row adjacent to. Here, the row adjacent to the first predicted sample may mean the upper row or the lower row of the first predicted sample. 2020/175915 1»（：1^1{2020/002754 At least one of the predicted samples used to induce the gradient may be adjacent to the first predicted sample and the other not adjacent to the first predicted sample. For example, based on a difference value between a second prediction sample positioned at the top or bottom of the first prediction sample and a third prediction sample spaced a predetermined distance in the X-axis direction from the second prediction sample, The horizontal direction gradient can be derived. The vertical gradient for the first prediction sample can be derived as a difference value between the prediction samples belonging to the column adjacent to the first prediction sample. Here, the column adjacent to the first prediction sample is derived from the first prediction sample. It can mean the left column or right row of the predicted sample. At least one of the predicted samples used to induce the vertical gradient of the first predicted sample is adjacent to the first predicted sample, and the other is adjacent to the first predicted sample. For example, based on a difference value between a fourth predicted sample positioned to the left or right of the first predicted sample and a fifth predicted sample spaced a predetermined distance from the fourth predicted sample in the y-axis direction. In the first example, it is possible to induce a vertical directional gradient with respect to the needle, where the predetermined distance may be a natural number such as 1, 2 or 3.

[536] Equation 18 is the horizontal direction gradient gradient H for the first predicted sample and

An example of inducing a direction gradient gradientV is shown.

[537] [Equation 18]

gradientH[x] [y] = (predSample [x+2] [y + 1]-predSample[x] \y+ l ^) ^' »shiftl gradieniV[x] [y] = (predSample [x+ 1] [y+ 2]-predSample [x+ 1] [y]) »shiftl

[538] In Equation 18, predSample represents the example, that is, the sample, and [Shi represents the x jug and y jug coordinates. shiftl represents a shifting parameter, which can have a predefined value in the encoder and decoder, or adapt the shifting parameter based on at least one of the size, shape, aspect ratio, or affine motion model of the current block. It can be decided as an enemy.

[539] When the gradient for the predicted sample is derived, the gradient and offset vector

Using, it is possible to derive an offset predicted value for the predicted sample. The offset predicted value can be derived based on a multiplication operation of a gradient and an offset vector. For example, Equation 19 shows an example of deriving the offset predicted value OffsetPred.

[54] [Equation 19]

OffsetPred[x] [y] = gradientH[x] [y] * offsetMV\x\ \y^S\ + gradientV * offsetA4V[x] [y] [1]

[541] If the offset predicted value is derived, the predicted offset value is added to the predicted sample,

The sample can be updated. Equation 20 shows an example of updating the predicted sample.

[542] [Equation 2

predSample [y] ⁼ predSample\x\ \y\ + OffsetPred[x\ [y^

[543] As another example, the predicted sample can be updated by adding an offset vector to the peripheral predicted sample. Here, the peripheral predicted sample is a sample positioned to the right of the predicted sample, a predicted sample. 2020/175915 1»（：1^1{2020/002754 The sample located at the bottom of the sample or the sample located at the bottom right of the prediction sample may include at least one of the samples. For example, Equation 21 is a peripheral prediction sample. This is an example of updating the prediction sample using.

[544] [Equation 21]

predSample [x] [y] ⁼ predSample [x+ 1] [y^ 1] + Offse tPred[x\ [y^

Information indicating whether to use an offset vector when performing motion compensation for the current block may be signaled through a bitstream. The information may be a 1-bit flag.

Alternatively, the use of the offset vector may be determined based on the size of the current block, the shape of the current block, or whether the affineseed vectors are the same. For example, a 4-parameter afine motion model is applied to the current block. If the first affine seed vector and the second affine seed vector are the same, motion compensation may be performed using the offset vector. Or, if a 6-parameter afine motion model is applied to the current block, the first affine seed vector Offset vector when both the affine seed vector, the second affine seed vector and the third affine seed vector are the same, or when two of the first affine seed vector, the second affine seed vector and the third affine seed vector are the same Motion compensation can be performed using.

[547]

[548] One prediction mode can be applied multiple times to the current block, or multiple prediction modes can be applied repeatedly. In this way, a prediction method using homogeneous or heterogeneous prediction modes is combined, eg, a mode (or Multi-hypothesis Prediction Mode). ).

[549] As for the combined prediction mode, the merge mode and the merge mode are combined, the inter prediction and the intra prediction are combined, the merge mode and the motion vector prediction mode are combined, and the motion vector prediction mode and the motion vector prediction mode are It may include at least one of a combined mode or a combined mode of merge mode and intra prediction.

[55 In this combined prediction mode, a first prediction block may be generated based on the first prediction mode, and a second prediction block may be generated based on the second prediction mode. Then, the first prediction block and the second prediction block may be generated. A third prediction block may be generated based on the weighted sum operation of. The third prediction block may be set as the last prediction block of the current block.

[551]

[552] In the above-described embodiments, the movement induced from the merge candidate under the merge mode

Various inter prediction techniques using information have been described. Specifically, inter prediction techniques based on the merge mode as follows are introduced through this specification.

[553] i) Regular merge mode: A method of performing motion compensation based on the motion vector derived from the merge candidate

[554] ii) Merge offset encoding mode: A method of modifying the motion vector derived from the merge candidate based on the offset vector, and performing motion compensation based on the modified motion vector. 2020/175915 1»（：1^1{2020/002754

[555] iii) Sub-block motion compensation mode: A method of deriving a sub-block motion vector based on the merge candidate and performing motion compensation on a sub-block basis.

[556] iv) Prediction unit partitioning-based encoding mode: A method of dividing the current block into a plurality of prediction units, and inducing the motion information of each prediction unit from different merge candidates.

[557] V) Combined prediction mode: A method of combining intra prediction and inter prediction (eg, merge mode)

[558] Information indicating that the inter prediction method based on the merge mode is allowed can be signaled through the bitstream. For example, the merge flag merge_flag indicates that at least one motion information of the current block is derived from the merge candidate. For example If the value of the syntax merge_flag is 1, it indicates that one of the methods is applied to the current block, e.g., an inter based on the merge mode described above. Either of the compensation mode, prediction unit partitioning-based encoding mode, or combined prediction mode can be applied to the current block, whereas when the value of the syntax merge_flag is 0, the inter prediction methods based on the merge mode described above are not applied to the current block. Represents

[559] Even if the value of the syntax merge_flag is 1, since various inter prediction methods based on the merge mode exist, additional information is required to determine the inter prediction method applied to the current block. For example, a subblock The syntax merge_subblock_flag to determine whether to apply the motion compensation mode, the syntax merge_offset_vector_flag or mmvd_flag to determine whether to apply the merge offset coding mode, triangle_partition_flag or merge_triangle_flag indicating whether to apply the prediction unit partitioning-based coding mode, or application of the merge_triangle_flag At least one of ciip_flag indicating whether or not may be additionally signaled. Or, whether the inter prediction method based on a specific merge mode is applied to the current block, instead of the flag, the size and shape of the current block, or the merge candidates included in the merge candidate list. It may be decided on the basis of at least one of the number of pieces.

[56] At this time, when other prediction methods other than the regular merge mode are not applied, the regular merge mode can be set not to be applied to the current block. However, in this case, it is determined whether or not the regular merge mode is applied to the current block. There is a problem with parsing many syntax elements, e.g. mmvd_flag indicating whether the merge offset encoding mode is applied, merge_subblock_flag indicating whether the subblock motion compensation is applied, i.e. merge_triangle_flag indicating whether the unit partitioning based encoding mode is applied, and/or If, after parsing all ciip_flag indicating whether the mode is applied or not, and then determining whether to apply the regular mode, there is a problem that many syntax elements must be parsed even though the regular merge mode is used more frequently than other modes. Determining whether or not regular merge mode is applied to the current block has several levels of parsing dependency. 2020/175915 1»（：1^1{2020/002754 Also, an index for specifying merge candidates under regular mode

There is a problem with high parsing dependency.

부호화하여시그날링할수 있다. [561] In order to solve the above problems, information indicating whether or not the regular merge mode has been used can be separately coded. Specifically, it indicates whether the regular merge mode is applied or not.

It can be encoded and signaled.

[562] Table 12 shows the syntax table including the syntax silver 111 Table 1_111 silver 6_: (¾ silver.

[563] [S12]

2020/175915 1»（：1^1{2020/002754

[4] The syntax is _11. Table 1 _{1 '_ 111} may be coded only if determined to be the inter mode based on the remaining yejeuk mode to ₆ ¾ _ :( eunneun current block is applied, that is, syntax

1 ₁₁ is ₆ _: (¾ is 1), the syntax is ₁₁ 1&]_ ₁₁₁ is ₆ _: (1& can be signaled through a bitstream.

[565] The syntax is the ₁₁ 1 Table _{1 '_ 111} _ :( ₆ ¾ autumn 1, the regular mode merge in the current block

If it is determined that regular merge mode is applied to the current block, one of the merge candidates is selected.

It can be signaled through the bitstream.

[566] When the syntax is 111_1116¾6_: (When ¾ is 0, the regular merge mode is

또는예측유닛파티셔닝기반부호화 모드의 적용여부를나타내는늘래그 1116 6_1；:1' 1¾16_：（¾은중적어도하나가 비트스트림을통해시그날링될수있다. If the syntax is 0, the flag indicating whether the merge offset coding mode is applied 11111 1_£ _§ , the Neulrag indicating whether the sub-block motion compensation mode is applied 1 high school 6 6_8111 10 0¾ is a combination example i.e. mode Flag indicating whether or not

Or, at least one of them can be signaled through the bitstream, indicating whether the prediction unit partitioning-based encoding mode is applied.

초하여 , 예측유닛파티셔닝 기반부호화모드의 적용여부를결정할수있다.일 예로,

참인것은현재블록에결합예측모드가적용되고,예측유닛 파티셔닝 기반부호화모드가적용되지 않음을나타낸다. cnp_nag7}거짓인 것은현재블록에 결합예측모드가적용되지 않고,예측유닛파티셔닝기반 부호화모드가적용될수있음을나타낸다. [567] Or, Neullag 1116 6_1::1' 1¾16_：（Omit the signaling of ¾ silver,

In addition, it is possible to determine whether to apply the prediction unit partitioning-based coding mode.

True indicates that the combined prediction mode is applied to the current block, and the prediction unit partitioning-based encoding mode is not applied. cnp_nag7} False indicates that the combined prediction mode is not applied to the current block, and that the prediction unit partitioning-based encoding mode can be applied.

[568] Or, a flag indicating whether the merge offset coding mode is applied or not. 11111 (1_典&)

Flag indicating whether sub-block motion compensation mode is applied or not

1116 6_8111 10 New 0¾ is a lag ()3_: (1& is) indicating whether or not the mode is applied. 2020/175915 1»（：1/10公020/002754

Flag indicating whether the subblock motion compensation mode is applied or not

그날링할수있다. In case, the regular merge is applied or not.

I can do that day.

[569] Instead of a flag indicating whether or not the regular merge mode is applied, a flag indicating whether at least one of the regular merge mode or the merge offset encoding mode is applied can be encoded and signaled. For example, the regular merge mode or the merge offset encoding can be performed. A flag 寒11101_111111¥(1_111 寒6_:^寒 indicating whether at least one of the modes is applied can be signaled through the bitstream.

[570] Table shows the syntax table including the secret tax for 111 1'_111111¥(1_1116 6_:^ dragon.

[571] [5.13]

2020/175915 1»（：1^1{2020/002754

[572] Syntax 111 table 1'_111111¥（1_1116 6_: (¾ is an inter-based interpolation based on the current block edge mode, eg, it can be coded only when the method is determined to be applied. That is, syntax 111 is 6_: In the case of 1, the syntax is 111&]'_111111¥（1_1116 6_：（1& is

It can be signaled through the bitstream.

[573] Syntax 6 indicates that the regular merge mode or merge offset encoding mode is applied to the current block when the value of 1'_111111¥（1_1116 6_:（¾ is 1, the current block is applied.

블록에 레귤러머지 모드가적용됨을나타낸다. There are 111 Table 1'_111111 ¥ (1_1116 6 _ :( ¾ autumn 1 ANN Wu, syntax 11111 (1_ _§ gachu indicates whether the remaining offset coding mode to be applied to the current block signaling. If the value of syntax 11111 （1_off& _§ is 1, it indicates that the merge offset encoding mode is applied to the current block,

Indicates that regular merge mode is applied to the block.

[574] Syntax 6 indicates that the regular merge mode and merge offset encoding mode are not applied to the current block when the value of 111 Table 1'_111111¥（1_1116 6_: (¾ is 0).

_ ¾16_：（¾은중적어도하나가비트스트림을 통해시그날링될수있다. When is 111_111111¥ (in the case of 1_1116, 10 0¾, which indicates whether the subblock motion compensation mode is applied or not, is a flag indicating whether or not the mode is applied, e.g., a combination indicating whether or not the unit partitioning-based coding mode is applied.

_ ¾16_: (At least one of the ¾ can be signaled through the bitstream.

[575] Or, it is possible to decide whether to apply the prediction unit partitioning-based coding mode based on [575] or, Neullag 1116 6_1::1' 1¾16_: (¾ is to omit the signaling, and based on (]3_: (¾). For example, being true indicates that the combined prediction mode is applied to the current block, and the prediction unit partitioning-based encoding mode is not applied. cnp_nag7} the lie 2020/175915 1»（：1^1{2020/002754 indicates that the combined prediction mode is not applied to the current block, and the prediction unit partitioning-based coding mode can be applied.

[576] Or, a flag indicating whether or not the subblock motion compensation mode is applied.

1116 6_8111 10 Shin 0¾ indicates whether or not the combined example, i.e. mode is applied.

Or a flag indicating whether to apply the prediction unit partitioning-based coding mode 1116 6_1;:1' 1¾16_: (¾ is at least one, and a flag indicating whether the regular merge mode or merge offset coding mode is applied is 111 Table 1'_111111¥ （11116 6_：（It is also possible to signal before ¾ is applied. For example, a flag 111 indicating whether subblock motion compensation mode is applied is 6_8111止10。] When 0¾ is 0, regular merge mode or Flag indicating whether the merge offset encoding mode is applied or not

Silver 111 votes 1'_111111¥（1_111 is 6_：（¾ can signal.

[577] Syntax 16 is 111 Table 1'_111111¥（1_1116 6_：（If the value of ¾ is 1, one of the remaining candidates

값보다작은경우,현재블록에 레귤러머지 모드또는머지오프셋부호화모드가적용될수있다.반면,

The merge index 111 for specifying is 6_:1 (signaled by encoding urine, and 111 is 6_:1 (1), the availability of the merge offset encoding mode may be determined based on the value of 1. For example,

If less than the value, regular merge mode or merge offset encoding mode may be applied to the current block.

If the value is greater than or equal to the threshold, the regular merge mode is applied to the current block, and the merge offset encoding mode is not applied.

[578] As an example, only when the value of at 111 is less than 2, the merge offset coding

이상인경우,머지오프셋부호화모드의 적용 여부를나타내는신택스 11111 （1_£ _§의부호화가생략되고,현재블록에 레귤러 머지모드가적용될수있다.즉, 의 값이 2보다작은경우에 한하여, 머지오프셋부호화모드가적용될수있다. Syntax 11111 indicating whether the mode is applied or not (1_ ^ _¾ can be encoded and signaled).

In this case, the encoding of syntax 11111 (1_£ _§) indicating whether the merge offset encoding mode is applied is omitted, and the regular merge mode can be applied to the current block. That is, only when the value of is less than 2, the merge offset encoding is performed. Mode can be applied.

[579] Syntax 11111 （1_ When _§ is 1, it indicates that the merge offset encoding mode is applied. When the merge offset encoding mode is applied, the merge index 111 can set the motion vector of the merge candidate pointing to 6_:1 (1) as the initial motion vector, i.e. 11111 （1_典& _§ 116 6_:1（Based on 1, the merge candidate used to derive the initial motion vector among the merge candidates can be specified.

[580] Also, if 111111¥（1_£1& is 1, then determine the offset vector.

出^此0!1_:1 (Urine can be signaled through the bitstream.

[581] As another example, when the merge index is 111 (Neullag indicating whether or not the regular merge mode is applied (for example, 111江1'_111 is 6_: (1&silver or 111江1'_111111¥（1_111 is 6_: You can configure the syntax table to be parsed earlier than (¾).

[58] Table 14 shows an example in which the parsing order of the merge index is parsed before the flag indicating whether the regular merge mode is applied. 2020/175915 1»（：1^1{2020/002754

[584] When it is determined that the merge mode-based motion prediction is performed in the current block, a merge index for specifying any one of the merge candidates may be signaled. That is, when the value of syntax 111 6_: (¾ is 1) , 111 6_:1（1 can be signaled to specify the future candidate.

이후, 레귤러 머지모드의 적용여부를나타내 [585] The flag indicating whether the regular merge mode is applied or not is the merge index 111 (which can be parsed after the urine is parsed. For example,

After that, it indicates whether the regular merge mode is applied.

As in the example in Table 14,

After signaling, the flag indicating whether the regular merge mode or merge offset encoding mode is applied can be set to be visually signaled 111 Table1'_1116¾6_: (¾).

특정되는 머지후보가현재블록의움직임 정보를유도하는데 이용될수있다. [586] When regular merge mode is applied, merge index

The specified merge candidate can be used to derive motion information of the current block.

는 머지 인덱스가문턱값보다작은경우에 한하여비트스트림을통해시그날링될 수있다. [587] When the merge offset encoding mode is applied, the merge candidate specified by the merge index 111 6_:1 (1 can be used to derive the initial motion vector of the current block. A flag indicating whether the merge offset encoding mode is applied or not.

Can be signaled through the bitstream only when the merge index is smaller than the threshold value.

[588] When it is determined that the regular merge mode is not applied (eg

Silver 111 Table 1'_1116 6_: (1&Silver or Silver 111 Table 1'_111111¥ (1_111 Silver 6_: (1&Silver is 0)), Neullag indicating whether sub-block motion compensation mode is applied or not 1116 6_8111 10010¾ Silver Visible Signaling Sub-block motion compensation mode is applied.

The merge candidate specified by can be used to derive the motion vector of the subblocks.

[589] When it is determined that the regular merge mode is not applied to the current block (for example, 111江1'_1116 6_: (1& is or 111江1'_111111¥ (1_111 is 6_: (1& is 0)) For example, the unit partitioning-based coding mode can be applied to the current block. The prediction unit partitioning-based coding mode is applied.

The candidate can be set as either the first partition or the second partition as a merge candidate. When the prediction unit partitioning-based coding mode is applied, the first partition or the second 2020/175915 1»（：1^1{2020/002754 A merge index has been added to specify the other one of the partitions.

Can be signaled.

[590] The flag regular_merge_flag or regular_mmvd_merge_flag indicating whether the above-described regular merge mode is applied can be signaled only when the size of the current block is smaller than the threshold. For example, when the size of the current block is smaller than 128x128, the regular merge is signaled. A flag indicating whether the mode is applied or not may be signaled. If the flag indicating whether the regular merge mode is applied or not is not signaled, it indicates that the regular merge mode is not applied to the current block. In this case, the prediction unit partitioning-based encoding mode can be applied to the current block.

[591]

[592] Intra prediction is based on the encoding/decryption completed restoration sample around the current block.

By using, the current block is predicted. At this time, for the intra prediction of the current block, a restoration sample before the in-loop filter is applied can be used.

[593] Intra prediction technique based on Matrix Intra prediction and surrounding restoration

It includes general intra prediction considering the direction of the sample. Information indicating the intra prediction technique of the current block can be signaled through the bitstream. The information may be a 1-bit flag. Or, the position of the current block, and Based on at least one of the size, shape, or intra prediction techniques of neighboring blocks, the intra prediction technique of the current block can be determined. For example, if the current block exists across picture boundaries, intra prediction based on a matrix in the current block Can be set to not apply.

[594] Intra prediction based on the matrix, pre-stored in the encoder and decoder.

This is a method of obtaining a prediction block of the current block based on a matrix product between the matrix and the restored sample around the current block. Information for specifying any one of a plurality of previously stored matrices can be signaled through the bitstream. have.

The decoder can determine a matrix for intra prediction of the current block based on the information and the size of the current block.

[595] General intra prediction, non-directional intra prediction mode or directional intra prediction

Based on the mode, it is a method of obtaining a prediction block for the current block.

[596]

[597] It is possible to induce the induced residual image by differentiating the predicted image from the original image. In this case, when the residual image is changed to the frequency domain, the subjective image quality of the image does not deteriorate significantly even if the high frequency components among the frequency components are removed. Accordingly, if the values of the high-frequency components are converted to small values or the values of the high-frequency components are set to 0, there is an effect that the compression efficiency can be increased without significant visual distortion. Reflecting the above characteristics, the residual image is displayed. The current block can be transformed to decompose into two-dimensional frequency components. The transformation will be performed using a transformation technique such as DCT (Discrete Cosine Transform) or DST (Discrete Sine Transform). 2020/175915 1»（：1^1{Wed 2020/002754.

[598] The transformation method can be determined in units of blocks. The transformation method can be determined based on at least one of the predictive encoding mode of the current block, the size of the current block, or the size of the current block. If it is encoded in the side mode and the size of the current block is smaller than NxN, the conversion can be performed using a conversion technique. On the other hand, when the above conditions are not satisfied, conversion technique 0 [can be performed using it. .

[599] 2D image conversion may not be performed for some blocks in the residual image.

have. 2D image conversion

When conversion skip is applied, quantization can be applied to residual values for which conversion is not performed.

용하여현재블록을변환한뒤 ,변환된현재블록을다시 변환할수있다.이때,。（그또는 에기초한변환을제 1변환이라정의하고, 제 1변환이적용된블록을다시변환하는것을제 2변환이라정의할수있다. [600]

After converting the current block by using the method, the converted current block can be converted again. At this time, the conversion based on that or is defined as the first conversion, and the second conversion is the conversion of the block to which the first conversion was applied. Can be defined.

[601] The first conversion can be performed using any one of a plurality of conversion core candidates.

For example, the first transformation may be performed using either DCT2 0018 or 1X17.

[602] Different conversion cores may be used for horizontal and vertical directions.

Information indicating the combination of the orientation conversion core and the vertical orientation conversion core

It can also be signaled through a bitstream.

[603] The units of the first transformation and the second transformation may be different. For example, the first transformation can be performed on an 8x8 block, and the second transformation can be performed on a 4x4 sub-block among the converted 8x8 blocks. At this time, it is also possible to set the transform coefficients of the remaining regions where the second transformation is not performed to 0.

Alternatively, it is also possible to perform a first transformation on a 4x4 block and a second transformation on an 8x8 area including the converted 4x4 block.

Information indicating whether or not the second transformation is performed may be signaled through the bitstream.

[606] Alternatively, on the basis of whether the horizontal direction conversion core and the vertical direction conversion core are the same, whether or not to perform the second conversion may be determined. For example, only when the horizontal direction conversion core and the vertical direction conversion core are the same, The second transformation may be performed. Alternatively, the second transformation may be performed only when the horizontal direction transformation core and the vertical direction transformation core are different.

[607] Alternatively, the second transformation may be allowed only when a transformation core defined as transformation in the horizontal direction and transformation in the vertical direction is used. For example, a 0012 transformation core for transformation in the horizontal direction and transformation in the vertical direction is used. If is used, a second transformation may be allowed.

[608] Alternatively, it is possible to determine whether to perform the second transformation based on the number of the thesis-zero transformation coefficients of the current block. For example, the thesis-zero transformation coefficient of the current block is less than the threshold value or In the case of 2020/175915 1» (：1^1{2020/002754, the second transform is disabled, and if the non-zero transform coefficient of the current block is greater than the threshold value, it can be set to use the second transform. It may be set to use the second transform only if the current block is encoded with intra prediction.

[609] The decoder can perform an inverse transform of the second transform (a second inverse transform), and perform an inverse transform of the first transform (the inverse first transform) on the result of the performance. The results of the second inverse transform and the first inverse transform are performed. , The residual signals for the current block can be acquired.

[610] By performing transformation and quantization in the encoder, the decoder can obtain a residual block through inverse quantization and inverse transformation. In the decoder, it is possible to obtain a restoration block for the current block by adding the prediction block and the residual block.

[611]

[612] When the restoration block of the current block is acquired, it is processed through in-loop filtering.

The loss of information that occurs during quantization and coding can be reduced. In-loop filters include deblocking filters, sample adaptive offset filters (SAO), or adaptive loop filters. ALF) may include at least one of.

[613]

[614] The application of the embodiments described centering on the decoding process or the encoding process to the encoding process or the decoding process is included in the category of the present invention. The embodiments described in a predetermined order are different from those described. Changing the order is also included in the category of the present invention.

[615] The above-described embodiment has been described on the basis of a series of steps or flow charts, but

The time series order of the invention is not limited, and may be performed simultaneously or in a different order as necessary. In addition, each of the components (e.g., units, modules, etc.) constituting the block diagram in the above-described embodiment is hardware It may be implemented as a device or software, or may be implemented as a single hardware device or software by combining a plurality of components. The above-described embodiments are implemented in the form of program instructions that can be executed through various computer components and are computer-readable. The recording medium can be recorded on a recording medium. The computer-readable recording medium may contain program commands, data files, data structures, etc. alone or in combination. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical recording media such as CD-ROMs and DVDs, magnetic-optical media such as floptical disks. media), and a hardware device specially configured to store and execute program instructions such as ROM, RAM, flash memory, etc. The hardware device may be configured to operate as one or more software modules to perform processing according to the present invention. It is the same as that of the weightlifting. 2020/175915 1»（：1^112020/002754 Industrial applicability

[616] The present invention can be applied to an electronic device that encodes/decodes an image.

Claims

2020/175915 1»（：1/10公020/002754 Claims [Claim 1] Whether the inter prediction based on the current block edge mode is applied Parsing the indicated first flag; If the first flag is true, parsing a second flag indicating whether a regular merge mode or a merge offset encoding mode is applied to the current block; And If the second flag is true, including the step of parsing a third flag indicating whether the merge offset encoding mode is applied to the current block, When the third flag is true, the merge offset encoding mode is applied to the current block, and when the third flag is false, the regular merge mode is applied to the current block. . [Claim 2] In paragraph 1, If the second flag is false, the video decoding method comprising the step of parsing a fourth flag indicating whether the combined prediction mode is applied to the current block. [Claim 3] In paragraph 2, The prediction unit partitioning-based encoding method is a video decoding method characterized by being applicable when the fourth flag is false. [Claim 4] In paragraph 1, The motion information of the current block is a candidate for the current block. Derived from the list, When the number of merge candidates derived from neighboring blocks of the current block is less than or equal to the threshold value, motion information candidates included in the motion information table are added to the merge candidate list as merge candidates. [Claim 5] In paragraph 4, When the current block is included in the merge processing area, the motion information table is not updated while the blocks included in the merge processing area are decoded. [Claim 6] In paragraph 4, When the current block is included in the merge processing area, based on the position of the current block in the merge processing area, it is characterized in that it is determined whether or not to update the motion information of the current block in the motion information table. ,Video decoding method. [Claim 7] In paragraph 6, If the current block is located in the lower right corner of the merge processing area, the 2020/175915 1»（：1^1{2020/002754 Video decoding method, characterized in that it is decided to update the motion information of the current block in the motion information table. [Claim 8] Whether the inter prediction based on the current block edge mode is applied Encoding the indicated first flag; If the first flag is true, encoding a second flag indicating whether a regular merge mode or a merge offset encoding mode is applied to the current block; And If the second flag is true, including the step of encoding a third flag indicating whether the merge offset encoding mode is applied to the current block, When the merge offset encoding mode is applied to the current block, the third flag is set to true, and when the regular merge mode is applied to the current block, the third flag is set to false. ,Image encoding method. [Claim 9] According to item 8, If the second flag is false, the image encoding method comprising the step of encoding a fourth flag indicating whether the combined prediction mode is applied to the current block. [Claim] According to claim 9, When the combined prediction mode is applied to the current block, the fourth flag is set to true, and when the prediction unit partitioning-based coding method is applied to the current block, the fourth flag is set to false. ,Image encoding method. [Claim 11] According to item 8, The motion information of the current block is derived from a merge candidate list of the current block, When the number of merge candidates derived from neighboring blocks of the current block is less than or equal to a threshold value, motion information candidates included in a motion information table are added to the merge candidate list as merge candidates. [Claim 12] According to claim 11, When the current block is included in the merge processing area, the motion information table is not updated while the blocks included in the merge processing area are decoded. [According to claim 11, When the current block is included in the merge processing area, it is determined whether to update the motion information of the current block in the motion information table based on the position of the current block in the merge processing area. 2020/175915 1»（：1^1{2020/002754 Characterized, image encoding method. [Claim 14] The method of claim 13, When the current block is located in the lower right inside the merge processing area, it is characterized in that it is determined to update the motion information of the current block in the motion information table, image encoding method.